offlineimap with more than two mailboxes?
Hi. Has anyone tried to use offlineimap to synchronize more than two mailboxes? I suppose it should just work, but if it does not, then I'd possibly lose some email. I was thinking of having one main server, A, and two sattelite boxes, B and C, syncing with the server. But I'd delete and move emails around in the satellite boxes, and expect the changes to reflect on the others. (I will *not* sync the satellite systems among themselves, as I guess this would probably confuse offlineimap) So -- did anyone try this before? Did it work? Thanks, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: converting file system
On Fri, Aug 03, 2007 at 02:35:09PM -0400, Douglas Allan Tutty wrote: So I doubt it. How you proceed depends on what mount point we're talking about. Hopefully, its not /. Anything else you can 'fix' by doing a backup, going single-user, unmount the partition, remake the filesystem, mount it, and restore the backup, then shutdown back to multi-user. Or use convertfs. (apt-cache show convertfs) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux on a Router
On Wed, Jul 11, 2007 at 02:00:11PM +1000, Alex Samad wrote: On Tue, Jul 10, 2007 at 09:18:43PM -0600, Nate Duehr wrote: OpenWRT being the more open of the options, is a good place to start, but might be a bit more manual than you want to set things up. I have been using openWRT for over 2 years and I have had not problems, my setup includes 2 uplinks and load balancing between the 2. They have 2 version 2.4 stream which they have come to end on and 2.6 there new release. Check out their wiki, it will have a table of hardware that work. I have been using OpenWRT also, and I think it's great. The package management system is very flexible, and the software is *very* stable. I tried the other Linux-on-router distros and none seemed to be as polished, robust and flexible as OpenWRT. Please check the webif^2 web interface also (http://www.x-wrt.org/). J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: FireFox slowness
On Thu, May 31, 2007 at 04:53:01PM -0400, Carl Fink wrote: Anyone else seeing absurdly slow responsiveness from iceweasel, like 30 seconds to close a tab? Yes, it happens to me also -- unfortunately, because I have gotten used to having several tabs open, and I keep opening and closing tabs all the time. Sometimes even switching tabs takes too long. :-( Or am I just using too many plugins? Hmm, I have eight plugins... Current Testing version. Unstable here. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Opening 300MB sent mail file
On Wed, May 30, 2007 at 09:45:57AM -0700, [EMAIL PROTECTED] wrote: What's a workable method of opening a 300MB file that I saved several years ago ? It's from Mozilla's email client, and it was an unorganized Sent Mail file. It's one huge concatenated set of emails. When I've tried to open it so far, I see 100% memory usage right away, but after long waits (half an hour) progress appears to slow to a crawl. When I most recently tried to open it with Mozilla-Thunderbird, I could read the beginning of the file (starting in 1998 !) but the screensaver came on before it opened all the way, and that seems to put the kibosh on finishing ... I have also tried OpenOffice, AbiWord, Mozilla-Firefox as well, with about the same result. My system has 500MB of memory and 10GB of free disk space that I can be sure of. Are there any debian app's which can handle this file ? Is it a mailbox file? Then you could use mb2md: apt-cache show mb2md Run it (with nice, ecause it will take a very long time), and you'll have a Maildir, which should be easier to access (it stores one message per file). J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Versioned /etc ?
Hi. So, I was trying to figure out a way to get automatic versioning for all files in /etc, and I wonder if someone tried that already. I would like to: 1. Not have to add files to a list when I add them to etc (like, I don't want to have to csv add any files); 2. Not have to manually check in (although I could get vim to automatically do that, I guess); 3. Be robust and not too complex. Simplicity is something I am willing to give up if nothing else works (but robustness is a must). What I could think of is: - Use version control. From simple rcs to full mercurial/monotone, this would be interesting, but this fails (1) and (2). Maybe a cronjob could automatically add unknown files, but I don't feel comfortable with that. (But I see lots of advantages in putting /etc in a distributed VCS, like automatic incremental backups, easy propagation to other similar hosts, etc); - Use a versioning filesystem, like CopyFS[0]. This is nice, but the only filesystems I could find with versioning are FUSE-based, and I'm not sure if it would be possible to get /etc mounted on FUSE. Maybe with initrd (anyone ever mounted root or /etc on FUSE?), but it fails (3). Does anyone have other ideas? Thanks, J. [0] http://n0x.org/copyfs -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bad cups printing after upgrade from sarge to etch
Hi. I had a sarge system running as a print server. After upgrading it to etch, I recreated the cups config files (the old ones didn't work) and the printers. The strange thing is... I have an Epson Stylus Color 850, and there are two drivers in the list given by Cups that would work with it: - CUPS+Gutenprint v5.0.0 - Foomatic/stc800ih.upp So... The Gutenprint driver doesn't seem to honor what I give in the configuration web page. It always renders with low resolution. I tried choosing all combinations of: Print Quality: Manual Control, High, Best Resolution: 720 DPI, 1440 x 720 DPI Highest Resolution on the web interface. And the stc800ih.upp prints with the resolution I choose, but always prints an EJL control sequence before printing the job: 284.4 @EJL I found a few references to this on the web, but nothing that actually helped (mostly Windows users complaining about it and being told to, er, remove the banner generated by the application). I didn't find anything in the bug tracking system that could help. Maybe there's something I forgot to check? Thanks a lot, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bad cups printing after upgrade from sarge to etch
On Tue, Apr 03, 2007 at 09:55:09PM +0200, Florian Kulzer wrote: I would try to change the quality setting by using the lpoptions command, both as your normal user and as root. Maybe this will turn up a clue. Is your normal user a member of the lpadmin group? No, but I can do that as root; The manpage of lpoptions has all the details; it should work something like this: $ lpoptions -l | grep -i quality Quality/Resolution, Quality, Ink Type, Media Type: *FromPrintoutMode 300DraftGrayscaleK 300GrayscaleK 600GrayscaleK $ lpoptions -o Quality=600GrayscaleK $ lpoptions -l | grep -i quality Quality/Resolution, Quality, Ink Type, Media Type: FromPrintoutMode 300DraftGrayscaleK 300GrayscaleK *600GrayscaleK OK, that works. There are two settings, Quality and Resolution. When I change them using lpoptions, then lpoptions gives back the changed values. (But the web interface doesn't). But the resolution is still the lowest. Then I tried setting with both lpoptions and the web interface. Didn't work either. So I tried something else: I used another driver, called Epson Stylus Color 800 Foomatic/stcolor (recommended) (en). It's not for the STC 850... But it works! (Although it won't print in 1440x720, but I don't need that -- 720 DPI is enough). Anyway, it seems taht the stc800ih PPD is broken. Nobody else had that problem? (The EJL control sequence being printed) Thanks, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: File encryption
On Sun, Jan 28, 2007 at 03:03:59PM +, Mark Crean wrote: If wonder if anyone's got experience or advice to share about a good way of using file encryption on Debian Etch? There seem to be a lot of different methods, but which one might suit the following: I only want to encrypt a single folder with personal stuff in it. Around If you want to encrypt one directory only, then you may want to take a look at encfs. (apt-cache show encfs). $ mkdir /path/to/encrypted/dir $ mkdir /path/to/clear/dir $ encfs /path/to/encrypted/dir /path/to/clear/dir (Now the encrypted dir is set up) $ cp myfile /path/to/encrypted/dir $ cd /path/to/encrypted/dir $ (edit myfile) $ fusermount -u /path/to/clear/dir And you'll see encypted files in /path/to/encrypted/dir; nothing on /path/to/clear/dir, since it was unmounted. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: help with C algorythm (find unique value in an array) could you please make changes
Hi. On Tue, Jan 23, 2007 at 06:01:34AM -0500, Roberto C. Sanchez wrote: On Mon, Jan 22, 2007 at 11:36:18PM -0500, Mike Polyakov wrote: Michael, Why not just use a std::setint here? Repeated inserts of the same value will be ignored. True, but that will use extra memory. Since pointers are iterators, this can be done on ordinary array in place without extra memory. Only thing is that unique() function modifies the original array. I have used back_insert_iterator with extra class to keep track of the number of unique elements, but now the code is not neat anymore: I think that std::set is the better approach. IIRC, the standard does not mandate how it is implemented, so it is very likely implemented as an array (at least for primitive types). gcc(g++) uses a red-black tree to represent sets: template class _Key, class _Compare, class _Alloc class multiset ... private: /// @if maint This turns a red-black tree into a [multi]set. @endif typedef typename _Alloc::template rebind_Key::other _Key_alloc_type; typedef _Rb_treekey_type, value_type, _Identityvalue_type, key_compare, _Key_alloc_type _Rep_type; /// @if maint The actual tree structure. @endif _Rep_type _M_t; ... The _M_t variable above is the multiset itself. This code is in the headers: bits/stl_set.h bits/stl_multiset.h J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian install hanging at yenta_socket module
On Thu, Jan 18, 2007 at 06:22:33AM -0800, Mitchell Verter wrote: The computer has just begun loading the CD and it seems to be hanging. This is what I see: Detecting hardware to find CD-ROM drives Loading module 'yenta_socket' for 'Toshiba America Info Systems ToPIC97' # Any ideas what this means? Any ideas what I should do? It happened with me with a Toshiba Tecra, and turning ACPI off fixed it (not sure if you have the same problem). Try booting with acpi=off as a kernel parameter. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What's your favourite FLOSS?
audio editor [ ] audio player [ cdcd, mpg123 ] cd-ripper [ crip ] Desktop Environment [ openbox ] DBMS [ postgresql ] development [ gcc, g++, bash, libboost*-dev ] disc burner [ wodim ] e-mail client [ mutt + offlineimap + fetchmail + mairix ] file manager [ bash + find + grep + sed ] finance [ ] ftp [ vsftpd ] image editor [ gimp ] image viewer [ gthumb ] instant messenger [ centericq ] mathematics [ bc, octave, GNU R ] misc utilities [ screen, wget, , grep, sed, encfs, rpl, xargs, less, more, ssh, ipe ] p2p [ ] package manager [ apt-get, aptitude ] pdf-reader [ xpdf ] spreadsheet [ ] tag editor [ ] terminal emulator [ multi-gnome-terminal ] * [0] text editor [ vim, tetex ] 3D animation [ ] video player [ xine, mplayer ] web browser [ firefox ] word-processor [ ] (unreleased) [ Debian kFreeBSD ] Not in the list, but should be: dns server [ ]* [1] game [ nethack ] smtp server [ postfix ] spam filter [ crm114 ] version control system [ monotone ] web server [ apache ] J. [0] Yes, the old multi-gnome-terminal. Tried several others, but none seems to be exactly what I need (and no more than that). [1] There is no DNS server that I really like. BIND is slow, bloated and insecure; MaraDNS is fast and secure, but lacks some features; I don't like the way djbdns is installed. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Window managers-which one?
On Tue, Oct 31, 2006 at 03:40:48PM +, Clive Menzies wrote: Since getting into Debian I've progressed down the scale (of bloat) from KDE to Xfce to Enlightenment to Fluxbox. I'm very happy now but guess I may get bored and try something else but fluxbox is lean mean but pretty functional. I went pretty much the same way, but then one day I thought fluxbox was kind of slow to draw menus etc... And I found openbox! It's fast, looks just like fluxbox, except that it doesn't have the extra fluff. :-) You may want to give it a try. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Window managers-which one?
On Tue, Oct 31, 2006 at 08:57:44AM -0800, Jason Dunsmore wrote: On 10/31/06, Jeronimo Pellegrini [EMAIL PROTECTED] wrote: On Tue, Oct 31, 2006 at 03:40:48PM +, Clive Menzies wrote: Since getting into Debian I've progressed down the scale (of bloat) from KDE to Xfce to Enlightenment to Fluxbox. I'm very happy now but guess I may get bored and try something else but fluxbox is lean mean but pretty functional. I went pretty much the same way, but then one day I thought fluxbox was kind of slow to draw menus etc... And I found openbox! It's fast, looks just like fluxbox, except that it doesn't have the extra fluff. :-) You may want to give it a try. I was a long time fluxbox user, but I didn't really like the task bar. Yes! Neither did I. And my openbox doesn't show one (it's optional). :-) I'd rather use something like WindowMaker, which manages windows more like a Mac. I used WindowMaker for a while, but it didn't work well with all programs. I finally found Enlightenment (pun intended). It's very stable and has just enough fluff, in the form of user feedback, so that it has a more solid feel than Fluxbox. I've found Enlightenment too bloated... But that's a matter of taste, so... :-) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Hot to get the original version of a conffile?
On Wed, Sep 13, 2006 at 11:07:59PM -0400, Kevin Mark wrote: On Wed, Sep 13, 2006 at 06:33:10PM -0300, Jeronimo Pellegrini wrote: Hi. I was looking at bug #247336, and wondering if it's possible at all to get the original conffile for an installed .deb. Is this information stored in the installed system somewhere? (Other than /var/cache/apt/archives, which does not necessarily have the .deb anymore) Hi Jeronimo, if you are looking for old version of .deb packages, they can (usually) be found on snapshot.debian.net. Hi Kevin. Actually, I wanted my backup script to find the original conffiles without having to connect to snapshot.debian.net or anywhere else... I suppose this is impossible? Is the conffile stored on my host, somewhere? Thanks, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Hot to get the original version of a conffile?
On Wed, Sep 13, 2006 at 11:08:59PM -0400, Wayne Topa wrote: Jeronimo Pellegrini([EMAIL PROTECTED]) is reported to have said: (I'd like to back up diffs of modified files, as opposed to a full dump of /etc). Not sure if I am understanding exactly what you want. I _think_ you are looking for *.dpkg-dist config files. If so then find /etc -iname \*.dpkg-dist should do it. Yes! Hey, that's nice. But only some conffiles have a .dpkg-dist version. How are they created? How can I be sure that all files taht I modified have a .dpkg-dist equivalent? I mean... If I install some package for the first time, I suppose there is no .dpkg-dist version of its config files. Thanks a lot! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Hot to get the original version of a conffile?
Hi. I was looking at bug #247336, and wondering if it's possible at all to get the original conffile for an installed .deb. Is this information stored in the installed system somewhere? (Other than /var/cache/apt/archives, which does not necessarily have the .deb anymore) Given a package, I know how to: - Get the list of conffiles - Get the md5 hash of each conffile (from /var/lib/dpkg/info) But I suppose the original conffiles are not stored? (I'd like to back up diffs of modified files, as opposed to a full dump of /etc). Thanks, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: wake-on-lan
On Tue, May 30, 2006 at 08:56:24PM +0200, Roel Schroeven wrote: Now I did investigate, and this is what I found out: given that the BIOS is configured correctly, I can remotely wake the box if I shut it down before it boots Linux (it works if I shut it down while the Grub menu is visible), but it doesn't work if I shut it down when Linux is running. I tried shutting it down via poweroff, shutdown -h now and pushing the power button, but all with the same effect: no response to wake-up packets. So, what can I do to enable the wake-on-lan feature again? I had a similar problem... Before shutting down, I need to run this command: # ethtool -s eth0 wol g Because after a reboot WOL is disabled. I included that line in my boot scripts. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Contact My Secretary
On Sat, May 06, 2006 at 02:28:48PM +0200, [EMAIL PROTECTED] wrote: Does anybody know how this scam works? I think this is what you're looking for: http://www.fbi.gov/majcases/fraud/fraudschemes.htm J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What is the point of sudo?
On Sun, Aug 28, 2005 at 06:58:37PM -0400, Michael Spang wrote: Ian wrote: I know it provides a fake root environment for work, but why would you want that? No, sudo allows root privileges on a per-user, per-command, and per-host basis. Yes -- but I would be careful when granting permissions with sudo. If the user will run any program that is able to edit (or even overwrite) files, then he'll be able to change /etc/passwd, or /etc/sudoers. If he will install packages, he will be able to configure packages too, possibly giving him more privileges than I'd want him to have... If he is able to use a front-end to some application that is able to edit files, then he's already root. I think this has been said here before, but since the original poster asked, and since I didn't see anyone mentioning this in the thread, I thought it would be good to mention it. Anyway -- it's a good idea to have sysadmins use sudo when you want to know who exactly did what, as someone else posted. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: diary with encryption
On Wed, Aug 17, 2005 at 10:46:26AM +0100, Oliver Elphick wrote: On Wed, Aug 17, 2005 at 06:57:53AM +0200, Laszlo Boszormenyi wrote: Hi, I am looking for an application which can store my notes in an _encrypted_ way. Preferably it is GTK+ or console based, and has some development to keep up with the current GTK+ version at least. You can use cfs to make a small encrypted filesystem out of a normal directory; then you can store your diary in the encrypted filesystem. That's what I would suggest, except that I'd use encfs instead of cfs. I had several problems with cfs, and encfs seems to be more stable and faster. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Request for window manager recommendations
On Sun, Jun 12, 2005 at 03:38:04PM -0700, [EMAIL PROTECTED] wrote: Not so keen on KDE/GNOME because as I understand they are somewhat CPU-intensive and take longer to load than the traditional WMs. A personal recommendation of your favourite window manager would be much appreciated. openbox -- it's very small, fast and configurable. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
usbmount only mounts pendrive as root:root?
Hello! So, after swithing one box from Fedora to Debian Sarge, there's onw thing users would probably like, but I don't know how to do: Fedora will mount pendrives automatically for you, with the permissions of whoever is on the console. I tried usbmount, but it seems to always mount as root. After the pendrive is plugged, this is how /media looks like: $ ll /media/ total 44K lrwxrwxrwx 1 root adm 4 2005-06-09 22:30 usb - usb0 drwxr-xr-x 2 root root 16K 1969-12-31 21:00 usb0 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb1 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb2 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb3 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb4 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb5 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb6 drwxr-xr-x 2 root adm 4.0K 2005-06-09 22:30 usb7 I had changed the group in *all* /media/usb* directories to adm, just to check if they'd go back to root:root when mounted, and indeed -- they do. So, is there a simple way to get the same behavior from Fedora in Sarge? Thanks a lot! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
GNOME/KDE in Sarge?
Hello. I have to install Linux on one box that I will admin remotely. A few users there will want to use GNOME (or maybe KDE, but it would be fine if I install just GNOME). I'll have to make a short trip to install the OS, and would like to do this as quickly as possible. So -- since I don't have a box here to try and test installing Sarge, my question is: how easy is it to configure X and GNOME? At least until Woody, it didn't just work out of the box, as it happens with Fedora, for example. Not that I think installing Fedora is a better idea, but I'd like to know how long it will take. Those who have installed Sarge -- what did you have to do to get X (and GNOME) running? Does it detect the hardware without problems? What is used to configure X? Did it work for you without need for manually tweaking /etc/X11/XF86Config-4 ? And did GNOME and gdm install nicely too? Thanks a lot! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: GNOME/KDE in Sarge?
On Wed, Jun 08, 2005 at 03:45:33PM +0200, David Mat wrote: Just installed sarge, the installer is great. Tthere is an option to autodetect your video hardware, so it'll work right away for some harware. Hm, that's what I wanted to know. With Woody, installing X didn't autodetect things properly (at least the last time I tried -- but I don't remember all details), and I had to go tweak the config manually. Thanks! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Is 64MB enough?
On Sat, May 21, 2005 at 10:00:54AM -0300, Rogério Brito wrote: On May 21 2005, Johan Kullstam wrote: John Moore [EMAIL PROTECTED] writes: Thanks but the old clunker's motherboard is not expandable to 256M :-( Star/Open-Office is not going to be pleasant. TeX, on the other hand, will run like a treat. Indeed. Let alone the typographical quality that you get with (La)TeX and that you don't get with other systems (if LaTeX is properly used, of course). I think it's not easy to use LaTeX so it will generate ugly output. I also don't use a desktop environment---I prefer to use a single window manager instead, and the solution that I found closer to my needs is fluxbox. I saw the light after experimenting with many window managers (like windowmaker, blackbox, oroboros etc) and some desktop environments (Gnome and KDE, mostly -- both inappropriate for a low resources machine). Fluxbox is small, functional, configurable, mostly useable with the keyboard (if you configure it correctly) and *fast*. Highly recommended. But it's getting more and more eye-candy stuff. Transparent menus, little icons along with menu entries... And it's getting big. It was quite slow on my box, so I switched to openbox -- and I can say it's FAST! BTW, looks at their sizes: -rwxr-xr-x 1 root root 1.2M 2005-04-19 13:41 /usr/bin/fluxbox -rwxr-xr-x 1 root root 200K 2005-02-21 19:59 /usr/bin/openbox I still do everything I used to do, without that extra 1Mb. I know the extra size doesn't necessarily mean it's slower, but it's more memory you'll be using, and if you're short on memory, it *will* slow down your system, no matter how good the code is. And fluxbox is very slow (at least it feels slow to me). I have 1.1 Mb RAM and I sometimes had to wait ten seconds or more just to open a menu. In the same environment, openbox opens menus instantly. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: md5sum input/output errors with large files
On Sat, May 14, 2005 at 03:30:12AM +, Pollywog wrote: When I do a md5sum check on large files (500MB or larger), I get this sort of error: [EMAIL PROTECTED]:~/RO$ md5sum RO_Beta_v3.2_Full.zip error processing RO_Beta_v3.2_Full.zip: failed in buffer_read(fd): mdfile: Input/output error Does anyone know what is going on? Are you using FUSE, LUFS, or any other usermode filesystem, or anything different than the standard ext3/reiser/xfs, etc? I had a problem with a FUSE filesystem that would show up as Input/output error. What does /var/log/syslog say? J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What BackUPS?
On Sat, May 14, 2005 at 10:55:34AM -0400, Colin wrote: Bill Day wrote: Well we had some rain and lightning and thunder and winds yesterday and i was asleep in my chair... lost all of my uptime becuase my old BackUPS was dead. Finally the question, what BackUPS do you use that Debian interacts with. The last one I had had no interaction with anything, it just beeped until it ran out of juice. I don't need a server side scale backup, just something that will give about 10-15 minutes for a proper shutdown. I use the apcupsd package to shutdown the computer before the battery runs out. It doesn't seem to work for me. After a while running and responding, the daemon seems to die, or stall. For example, in the first time I use apcaccess (right after booting) it works. But after some time, it stalls and doesn't respond. Trying to restart apcupsd also doesn't work. I run sid, and I've been compiling kernels from kernel.org (I believe this may be the reason). What are you running (kernel and version of Debian)? J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 'Virtual Private Servers' - Advice, recollections and recommendations requested
On Mon, Jan 10, 2005 at 05:57:56AM +1100, Sam Watkins wrote: I also use Linode, Linode is great, I have no complaints! They have a forum (quite like fastmail.fm's one) where you can get help from other users and the developer of Linode, and they use a special kernel hack to prevent other people on the same box from monopolising the disk IO and slowing the machine to a halt. The web interface is good. I think I am right in saying that Linode is widely regarded as the best UML virtual server provider (at least among Linode users!). I haven't tried others... But I know at least one that seems to offer more disk space and bandwidth for the same price: http://rimuhosting.com/order/startorder.jsp They also offer backups, which Linode doesn't, as far as I know. Wasn't there a page somewhere with a list of companies offering UML hosting? I think I saw it once, but I never found it again. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 'Virtual Private Servers' - Advice, recollections and recommendations requested
On Fri, Jan 07, 2005 at 02:04:34PM -0500, Roberto Sanchez wrote: The hosting provider I mention above offers 600 MB space and 25 GB/month (for the $70 plan) or 800 MB space and 35 GB/month (for the $100 plan). That should be plenty for most any organization. Since I wanted root access, I got it (with Woody preinstalled) at linode.com (they use UML). For $20/mo, you get 25Gb transfer, 64Mb RAM, and 3Gb disk space. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Script similar to Redhat chkconfig
On Thu, Dec 30, 2004 at 05:33:28PM +0100, Andrea Vettorello wrote: On Thu, 30 Dec 2004 11:15:47 -0500, Ian Meyer [EMAIL PROTECTED] wrote: Does something like this exist for Debian? I know the update-rc.d is there to add/remove services to rcX.d but chkconfig will list the services and what their runlevels are and whether or not they are started/stopped. Just curious, is all. Not present. There's a bug filed (#227604) in the Debian BTS. About the services, the default is to have them all enabled on all runlevels, with 1 runlevel 6... =) What about sysvconfig? BTW, #227604 has been closed... J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Is Linux Unix?
There's another problem with the above C++ code: If ny and nz
aren't constant, you can't write
double s[ny][nz];
Instead, either you allocate the array in two stages:
double** s = new (double*)[ny];
for (j=0; j ny; ++j) s[j] = new double[nz];
Actually, the first version works -- and both C and C++ (tested here with
gcc -- not sure it it became a standard or not)) will dynamically allocate
memory for you. Try this:
#include iostream
using namespace std;
int main() {
int size1, size2, i, j;
cin size1;
cin size2;
double vec[size1][size2];
for (i=0; isize1; i++)
for (j=0; jsize2; j++)
vec[i][j] = i+j;
cout endl;
for (i=0; isize1; i++) {
for (j=0; jsize2; j++)
cout vec[i][j] ;
cout endl;
}
}
And enter something like 1000 and 1000 when asked for the array dimensions.
It compiles, runs, and does not segfault! :-)
The C version:
#includestdio.h
int main() {
int size1, size2, i, j;
scanf(%d, size1);
scanf(%d, size2);
double vec[size1][size2];
for (i=0; isize1; i++)
for (j=0; jsize2; j++)
vec[i][j] = i+j;
puts();
for (i=0; isize1; i++) {
for (j=0; jsize2; j++)
printf(%f ,vec[i][j]);
puts();
}
}
J.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: yay! I can print!
On Sun, Jul 25, 2004 at 12:40:05PM -0400, Jason Rennie wrote: mpage? Or enscript? J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling maildrop with mysql enabled fails on sarge
./configure: line 1: mysql_config: command not found configure: error: Unable to run mysql_config configure: error: /bin/sh './configure' failed for maildrop I am no developer-person so I am lost here... And I really need the mysql-support into maildrop :(. Install libmysqlclient-dev (or something similar -- in woody, it's libmysqlclient10-dev) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling maildrop with mysql enabled fails on sarge
debian/maildrop/usr/bin/maildrop: error while loading shared libraries: libfakeroot.so.0: cannot open shared object file: No such file or directory dpkg-shlibdeps: failure: ldd on `debian/maildrop/usr/bin/maildrop' gave error exit status 1 dh_shlibdeps: command returned error code 256 What am I missing here? I think you'll find the answer here... http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=253436 J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling maildrop with mysql enabled fails on sarge
Jul 11 20:07:09 srv1 postfix/pipe[29843]: 359E431346: to=[EMAIL PROTECTED], relay=maildrop, delay=526, status=deferred (temporary failure. Command output: maildrop: signal 0x0B ) It is a long road to getting this to work I don't know what signal 0x0B means, but maildrop doesn't seem to be able to deliver to the mailbox where you told it to do. Check these: - Is the path to the mailbox correct? How does it appear in the database, and what is the path to your maildirs? - Permissions and ownership in the mail dir? These should be: drwx--9 vmailvmail J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: migration of Maildir to new system. Also advice wanted on performance
Thanx for the reply Jim, but please post to the list so everyone can read :). Anyway, is there anyone who can answer my other questions? How big is the performance gap between Maildir / Cyrus? Is it worth it? I don't know about that, but I didn't install Cyrus because upsteam considers the 1.x branch not safe to use. From http://asg.web.cmu.edu/cyrus/download/ : | The 1.x series of Cyrus IMAPd has been deprecated. It should be | considered unsafe to use. And the 1.x branch is the only one in Woody... And backports of Cyrus usually bring lots of backported dependencies which I didn't want. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: migration of Maildir to new system. Also advice wanted on performance
On the other hand, I'd prefer to stick with my current stuff since I now it. So this makes it Postfix + Courier again. One last consideration for me is a nice web-based admin tool for virtual domains and virtual users. I would like to store the info in LDAP or MySQL. I had quite some issues with OpenLDAP in the past so I guess MySQL is preferred for me. I did see a lot of web based stuff, but not for Postfix. Then you seem to be looking for PostfixAdmin: http://high5.net/postfixadmin/ J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: RH and Debian brothers now?
On Thu, Sep 25, 2003 at 06:11:53PM +0200, Sven Hoexter wrote: On Thu, Sep 25, 2003 at 03:43:00AM -0700, Paul Johnson wrote: 1) apt-rpm is a piece of shit compared to apt-get due to RPM stupidity. Aehm the only difference I can find from user standpoint is that apt-rpm is slower cause auf the rpmlib. And sometimes you have to maunally remove a lock file, or it just won't work. And it won't tell you why it's not working (it stalls, and you have to guess what's going on). J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Help detecting possible intrusion
Hello. Today I found one of my servers (Woody on an uml kernel) was down. It's in another country, but I can admin it remotely. I rebooted it (uml lets you do that), and found a couple of strange things. - AIDE tells me all /dev and some tty devices were created right before the server crashed: Example: changed:/dev changed:/dev/ttyp0 changed:/dev/ttyp1 changed:/dev/ttyp2 changed:/dev/ttyp3 changed:/dev/ttyp4 changed:/dev/ttyp5 changed:/dev/ttyp6 changed:/dev/ttyp7 ... Directory: /dev Ctime: 2003-09-01 16:48:44 , 2003-09-25 18:53:30 File: /dev/ttyy3 Ctime: 2003-09-01 16:48:42 , 2003-09-25 18:53:24 What does that mean? - We run bsd-ftpd, and I last tells me: ftp ftp Thu Sep 25 18:57 - 18:57 (00:00) ftp ftp Thu Sep 25 18:57 - 18:57 (00:00) This was right after reboot (if not during it). But the ftpd logs say nothing about this guy. Does bsd-ftpd only log transferred files, or does it also log logins? There was nothing in kern.log and syslog showing why it crashed. The company hosting this says their UPS and backup generators would hold the system up and running in case of a power outage... chkrootkit finds nothing strange (I rsync'ed a new version to the server, didn't trust the one there). Does that sound too bad? I'm particularly worried about the /dev/ ctimes changed before the crash. Any ideas? Thanks! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
How can I find out the SA (v2.55) scores of all email? I think it's in /usr/share/spamassassin, in the .cf files. I ask that because I'd like to see how it scores spam that it thinks is ham. Once I know that, I know which knobs to tweak. Change the scores in /etc/spamassassin/local.cf J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
On Tue, Sep 23, 2003 at 06:10:42AM -0500, Ron Johnson wrote: Ok, clarification: I'm not asking for the score parameters; I'm asking for the score of each individual email. Isn't it in the headers? My configuration does nothing but to put names in whitelists/blacklists, and I always get the scores on the headers: X-Spam-Status: No, hits=-7.2 required=4.0 tests=BAYES_01,EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT, REFERENCES,REPLY_WITH_QUOTES,SIGNATURE_LONG_SPARSE,X_LOOP, X_MAILING_LIST autolearn=ham version=2.55 (This is from your message) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
On Tue, Sep 23, 2003 at 01:24:52PM +0200, Robert Vollmert wrote: easiest seems to be to pipe the message to 'spamassassin -t'. Depending on your mailclient, of course. I'm sure it's also possible to make spamassassin insert its report header into every mail it checks. I always pipe it through spamc, and the report is added to the headers, even when it's ham. Doesn't that work for you, too? J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: en_US.UTF-8 compose bug fixed, but whom should i send it to ?
searching a little bit i found out the the compositions were defined but in the wrong (i.e. different than what i'm used to) order of key strokes. So I just patched the file /usr/X11R6/lib/X11/locale/en_US.UTF-8/Compose to include any order (for all key compositions). i'd like to send the fix to the proper debian maintainer(s). How do I find this out ? dpkg -S /usr/X11R6/lib/X11/locale/en_US.UTF-8/Compose xlibs: /usr/X11R6/lib/X11/locale/en_US.UTF-8/Compose So, check the bugs in xlibs: http://bugs.debian.org/xlibs And if this bug isn't there, report a new one and include your fix. (I'd suggest reportbug for that) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
It does, actually. I was thinking of the more detailed report that comes in the body of spam messages, and which lists each matched test's score and description. Ah, I see. I used to check each score in /usr/share/spamassassin, but I agree -- piping it through spamassassin -t (as you suggested) is easier. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
On Tue, Sep 23, 2003 at 08:48:44AM -0500, Ron Johnson wrote: On Tue, 2003-09-23 at 06:28, Jeronimo Pellegrini wrote: X-Spam-Status: No, hits=-7.2 required=4.0 tests=BAYES_01,EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT, REFERENCES,REPLY_WITH_QUOTES,SIGNATURE_LONG_SPARSE,X_LOOP, X_MAILING_LIST autolearn=ham version=2.55 (This is from your message) Nope. I meant it is how (my) spamassassin classified the message. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
On Tue, Sep 23, 2003 at 09:24:51AM -0500, Ron Johnson wrote: On Tue, 2003-09-23 at 07:44, Richard Humphrey wrote: Add this to your local.cf file and each email will contain the results of the tests, so you can compare ham vs spam etc. always_add_report 1 Ok, I added these to /etc/spamassassin/local.cf, and restarted SA, but still don't have X-Spam-Status in new emails always_add_headers 1 always_add_report 1 fold_headers 1 It's funny... I don't have those, and the headers do have the (terse) report. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SpamAssassin and false-negative spam
On Tue, Sep 23, 2003 at 09:32:14AM -0500, Ron Johnson wrote: Yes, I know. I mean that I don't have anything like that in any of my mails, neither ham nor spam. (But the spams get quarantined and I get a SPAM FROM email from amavisd-new, so that doesn't bother me.) Wait... You're calling spamassassin from amavis? So the place to configure it is in the amavis config file. But I think it's not very flexible... In one server, I had to do that, and I couldn't get the report for ham (but I did get it for spam). J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Anti-Spam ideas for usenet/list harvested email addresses
On Tue, Sep 23, 2003 at 01:16:38PM -0600, Jacob Anawalt wrote: I've already mentioned the web authorization idea and the rotate your email address on some schedule ideas in another thread. I've even seen a web site go so far as to use a .js file function to put together the email address from a bunch of fragments when you click the mailto link. That would take more work to parse, but it is still possible by having an email grabbing webbot that can run javascript. That would also break for people who use non-Javascript enabled browsers. Another though I've had on the mailing list issues (besides wondering why I'm trying to make mail act like a news client with threads and looking for a 'watch thread' capable client) is if I had an email address to use on mailing lists that only accepted email from the list servers I was on and reject all others I should only get the spam that relayed through the list. Interesting. But managing that would require some energy from you... The mail server would need to have access to my personal list of acceptable email addresses so it could give a 550 with the appropriate extended SMTP code for unauthorized/security and an appropriate error message after the HELO and MAIL FROM and RCPT TO: have been given. It should only do this for mail accounts that have entries in the safe list. If your list is empty, all email is valid. If you have one or more entries, only those ones can send you email. Some ideas for rules to accept or reject the email may include: If HELO does not match a reverse DNS lookup and doesn't match the domain of RCPT TO: or to a user specified value then the mail is rejected. Blocks big ISPs... I've found two already. One of them is movistar.com. Can't remember the other Also, probably breaks small businesses who use DSL and can't use their ISPs smarthosts (see the recent thread, OT: Martin Krafft - mail bouncing. A looser match would be just on the HELO name where the name given is some md5hash of the user's email address and some value noted on the mailing list. People start getting spammed, the list admin changes the key used to generate the name value and people go to the web to see what it has been changed to. If I use taht, I'll have to keep changing the key every now and then. Spam is bad not only because it takes a lot of bandwidth, but also because it's not convenient. Challenge-response solution can be as inconvenient as spam itself, for example. And I think the same would work for this solution... I'm sure there are other better ideas to be had along the lines of how to quickly identify that the sending server is who they say they are and look up a safe list to see if the user accepts email from that server. Make the list server PGP-sign the messages, maybe? You install the list server key once, and never worry about it again? Compare this to the dog chasing cars method of inventing a new filter rule that looks through the MIME data to decide if this is the latest worm you don't want or the kissing picture that you do. Sure it's cool to be a geek and figure out the rules. If you like doing this, do it. Maybe spam isn't a cost to you but a benifit if you consider your enjoyment at solving each filter puzzle. I think that's why I like finding bugs, to help find and solve puzzles. On the other hand this method of filtering is more expensive in every measure I can think of except the freedom of allowing anyone to email you anytime. You spend time thinking up rules, writing rules and testing rules. The rules are applied after you have accepted the bandwidth of the transfer. Running the rules takes CPU time and possibly more bandwidth as you do RBL DNS or Razor and storing the email takes disk space. I agree. But then I think any technical solution has the same problem. The real solution would be making spammers not want to spam (so we don't have to block them). You'd need to understand the intricacies of their business, and so something that makes them give up. A very naïve thing would be to start doing statistical research, asking people how they feel when they get spam, and make that get to the clients of these spammers. But as I said, this is naïve, and assumes that we know how that business works. (I don't think I know that) But something along those lines will have to work, someday -- I hope! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Anti-Spam ideas for usenet/list harvested email addresses
But my goal was to reduce the spam I get that is harvested from mailing lists. If someone wants to subscribe to a mailing list that doesn't do reverse dns, then there needs to be authentication before DATA on some other bit of information. I could still get posts from the guy in Brazil or the guy using SMTP off of his cable modem DHCP'd address because they would be mailing the list, not me. The list is mailing me. Hm, that's right, indeed. Make the list server PGP-sign the messages, maybe? You install the list server key once, and never worry about it again? If some small PGP/GPG data could be sent as part of a new EHLO syntax command then OK, otherwise I'm in the DATA section again. It would have to be a standard before I'd use that. You want to reject the mail before it's queued. I like the idea, but that's more difficult to implement... I wonder how many MTAs would let you do this: - set up a mail for lists only - set up terribly-aggressive blocking with DNSBLs and other things (like requiring the reverse DNS), *only for that address*. Other addresses would not go through such restrictive tests. The latest churn on debian-user about Spam hasn't been UCE spam. It's been worm spam. I don't know anyone personally who likes to recieve WORM/Virus code in their inbox but it persists. I don't see a near-term solution for convincing the individuals who write this code. Right, I forgot about that. Anyway... Blocking servers wouldn't help in the case of viruses, I think. Ordinary people get viruses, and the mail is sent through their (probably correctly configured) smarthost. Maybe something like Postfix header_checks? But that would also require some work :-( J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 04:42:21PM +0200, martin f krafft wrote: We have taken the discussion up in private. The problem is in fact the dynamic IP of the dialup, which I filter using the dynablock RBL. It just happens that these RBL filter 65% of all my spam before it hits the content filters. At peak 150k mails/day this makes the difference between a usable system and one that is DoS'ed. I've managed to block a lot of spam (didn't measure, but my guess is that I blocked 70%), using only lists that do not block dialup: reject_rbl_client relays.visi.com, reject_rbl_client relays.ordb.org, reject_rbl_client sbl.spamhaus.org, reject_rbl_client proxies.relays.monkeys.com, reject_rbl_client opm.blitzed.org, reject_rbl_client cbl.abuseat.org You may try these... They should leave a manageable amount of mail to spamassassin, or wharever else you use after the blocklists. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 12:32:38PM -0300, Christoph Simon wrote: Unfortunately, there are many private victims for false positives of RBL-like lists, according to them, mostly due to the lack of response from our ISPs. As a matter of fact, I do have a fixed IP but that is taken out of a range of IPs mostly used for dynamic assignement. To make it worse, the ISP denies delegation of the reverse resolution. I know lots of small businesses in Brazil that are in the exact same situation... The problem is the administration of these RBL lists, which either tell you that any kind of communication with them will be published on usenet (including valid email addresses), as they presuppose that everybody in their list _is_ a spammer, or just don't give any chance to contact them. Although I can't contribute anything constructive to the above discussion, I do want to use this context to apeal these list's users, trying to convince their maintainers, that false positives do hurt people in many ways and that not being able to tell them, does'nt really help. I think it depends on how you choose the lists you're going to use. For example, ordb only lists open relays, and inclusion/exclusion is automated. There is also relays.visi.com, which seems to be very conservative. I've had luck with proxies.relays.monkeys.com, too (no false positives at all), but I can't say they're conservative. But I decided not to block dialup... I'd be blocking some mail servers I've configured myself! :-) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 05:49:21PM +0200, martin f krafft wrote: If your ISP is being a bitch about it, then switch! Otherwise just relay via their SMTP smarthost and the problem is solved. Unfortunately, we don't have a lot of options. And the SMTP smarthost is veeery unreliable. Quite a mess. :-( J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 05:55:20PM +0200, martin f krafft wrote: I already use all of these (plus ordb.org), but most of the spam (and most of the virus crap) is filtered by dynablock. Did you try putting dynablock at the end of the list, so as to check if some dynablock rejects wouldn't be caught by the others first? I don't see why people don't relay via their ISPs. Is there one good reason? Sure: the smarthost doesn't work well, the ISP doesn't care, and for example, in the state of São Paulo, you either get a wireless connection (and that depends on where you are, and you need to be in a tall building), or you get DSL from Telefonica (ask anyone from Brazil abou them...) -- and what they call a business DSL connection doesn't even stay up 24/7. They disconnect you periodically so you have to authenticate again. Aaargh! We really should have more options. :-( There are other problems too, like, we have to pay the DSL provider, plus another ISP (dumb politics), the DSL provider doesn't offer SMTP hosts. Some people say they don't need anything but the DSL provider, and then they send email from their own servers... Anyway -- the situation is a mess, but the point is, quite some legitimate business are blocked by DULs. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 05:23:48PM -0500, John Hasler wrote: Nothing in particular, in general. For example, there is only one drugstore in my village. The owner is entirely free to set his prices however he wishes. Same goes for my ISP. I think the point is that in Brazil you can't start offering DSL service. The monopoly is sort of enforced by a regulating agency. (The original idea was not that, but this is what happens in practice) So, they do whatever they want with their prices, and nobody can get an alternative. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Martin Krafft - mail bouncing
On Sun, Sep 21, 2003 at 07:04:07PM -0500, John Hasler wrote: Jeronimo Pellegrini writes: I think the point is that in Brazil you can't start offering DSL service. The monopoly is sort of enforced by a regulating agency. And thus we have an example of the evils of regulation, not of the evils of monopoly. Yes, yes... My point is dial-up lists block legitimate email; may be because of regulation, or whatever else. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: make-kpkg ( line 2: syntax error: unexpected end of file)
fakeroot make-kpkg --append-to-version=.20030812 kernel_image Which is (for the 2.6 try): lots and lots of output.. . and finally make[1]: Leaving directory `/usr/src/kernel-source-2.6.0-test2' echo done stamp-build /usr/bin/make -f /usr/share/kernel-package/rules real_stamp_image make[1]: Entering directory `/usr/src/kernel-source-2.6.0-test2' /bin/sh: -c: line 2: syntax error: unexpected end of file What version oa kernel-package? See http://bugs.debian.org/204856 (It's fixed in the latest version -- I've compiled test3 here without problems) J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: make-kpkg ( line 2: syntax error: unexpected end of file)
The interfaces have also changed, meaning that the nvidia driver you got from nvidia will not work with newer kernels without some hacking. Just google around and check nvnews.com, as there are plenty of kernel hackers out there on the bleeding edge that have gotten theirs to work and then supplied patches. I've heard the best one is this: http://minion.de/ (the guy seem to have worked for NVidia). J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Apache character encoding problems [ FIXED; thanks ]
Thanks to the guys who answered in private (but do answer to the list next time!) I had to comment out this line: AddDefaultCharset on in httpd.conf, so now Apache will not force ISO-8851-1. I understand there's a security issue involved (http://www.cert.org/advisories/CA-2000-02.html), but we do trust the people who will be writing the webpages enough. Again, thanks. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Apache character encoding problems
Hello. I'm trying to migrate one web server to one Debian box, and almost everything works... Except for one little problem. This server hosts sites in different languages, and it seems that only the ISO-8859-1 languages are being shown correctly (Spahish, English, Portuguese, Italian, etc). All others show garbage (Chinese, Bulgarian, Greek, etc). The server is hosted at a company that offers user-mode Linux, and we have chosen Woody for that box. The Apache server is the one in woody (1.3.26-0woody3). I spent the last two days searching and reading documentation, but didn't find anything (tried Google and the Debian BTS). Did anyone else experience such a problem before? Is it our Apache configuration or could it be something else? (We're migrating from Red Had to Debian) The same problem occurred when I tried switching to the unnofficial debs from http://people.debian.org/~nobse/debian/woody/backported/apache2/ Apache packages installed: apache apache-common apache-doc libapache-mod-perl libapache-mod-gzip Unnofficial: apache2-common apache2-mpm-prefork libdb4.1 libapr0 db4.1-util (If it helps, I can post a test URL) Thanks a lot, J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Plagiarism Monitor S/W
On Mon, Feb 10, 2003 at 06:15:15AM +0800, [EMAIL PROTECTED] wrote: On Sat, 8 Feb 2003 00:04:12 -0200, Jeronimo Pellegrini wrote: I know one person who used to teach Compiler Construction. He had a program that parsed the student's compiler, built a syntax tree, and used an algorithm to compare trees. Anything over a certain level would trigger an alarm, *and then he'd check it himself*. Such a tool wouldn't ring the bell if a student had shared a few pieces of code with another -- but it'll certainly catch the guy whose work is mostly copied from someone else (no matter how much he changes in variable names or comments!) Probably (the instructor) didn't like Open Source. He uses Linux, and advocates it. :-) But the tests were supposed to be done individually, and two syntax trees that are too similar usually mean someone cheated. (Copying and changing variables and comments is not open source development...) -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: Plagiarism Monitor S/W
On Fri, Feb 07, 2003 at 01:34:31PM -0500, David Z Maze wrote: So, having had some experience doing this: your class has TA's, right? And they review the things students turn in? When I've been a TA, this has caught the more gratuitous cases of cheating; having a class policy that code sharing is absolutely verboten esssentially forbids students from working together at all, which is counterproductive. A sufficiently aggressive tool will result in lots of false positives, too, which isn't helpful. Finally, if you do decide to go after students, please apply some discretion, and don't assume your tool infallibly detects the students' intentions... I know one person who used to teach Compiler Construction. He had a program that parsed the student's compiler, built a syntax tree, and used an algorithm to compare trees. Anything over a certain level would trigger an alarm, *and then he'd check it himself*. Such a tool wouldn't ring the bell if a student had shared a few pieces of code with another -- but it'll certainly catch the guy whose work is mostly copied from someone else (no matter how much he changes in variable names or comments!) J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: finite-state automata in LaTeX
On Wed, Feb 05, 2003 at 01:20:44PM -0500, Xavier Barnabe-Theriault wrote:
Exporting in ps/latex creates two files. One is the pstex which is an
eps file with another extension, containing all non-special text stuff,
as you've seen it ! The other, pstex_t, is a latex file
containing latex commands to display and interpret special text. You
must input this file like
\resizebox{10cm}{!}{\input{../seminaire/images/flux-analyse.pstex_t}}
Ah, that works! And the result is quite nice.
Thank you! (And thanks to the other guys who answered!)
J.
--
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: finite-state automata in LaTeX
On Sun, Jan 26, 2003 at 09:16:29PM -0500, Hubert Chan wrote:
Actually, if you export as Combined PS/LaTeX, it can used LaTeX to
typeset the text part of the figure. And if you set the special flag
in your text (click on the text tool. Then, at the bottom of the
screen, click on the Text Flags button. Click on the button next to
Special Flag and select Special), you can use TeX commands
(e.g. math mode) in your labels.
I tried it here and it didn't work... Exporting works OK (but you need
the transfig package installed); I used:
\usepackage{epic,eepic}
\usepackage{psfrag}
(...)
\includegraphics{Automaton.pstex}
Ran latex, but when I tried to see it with xdvi, the text was missing in
the picture. (It looks the same if I try to directly open Automaton.eps
with gv).
I see there is a file called Automaton.pstex.aux (which means latex
processed it), and a Automaton.pstex_t file, with the latex code to
include the text.
Is there something else that needs to be done?
Thanks,
J.
--
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ISP does not 'support' Linux
I'll make some comments, in the hope that they'll help. On Fri, Nov 29, 2002 at 10:35:01AM +, Chris Lale wrote: Here's an idea arising from the 'Non-Linux-aware ISP: please spoon feed' thread. How many ISP's helplines say 'we do not support Linux'? Most ISP's seem to have a webpage with connection instructions for Windows users. Why not instructions for Linux? I remember they used to say (at least here in Brazil) that thre are too many different ways to connect using Linux, and too many different Linux distributions. That does make sense, unfortunately: the help desk people are usually not very technically-oriented; they are just people who can follow a script (it's cheaper). It'd be difficult to make them learn all different problems a non-knowlegeable Linux user may have when connecting. Of course, an experienced Debian admin would probably not call them except to get essential information (DNS servers, gateway, password, etc), but what about the guy who's beginning with Linux (or some similar situation)? What if the guy bought some book (like Linux Unleashed) that tells him do do things in a way that's different from what the ISP would recommend? It's not difficult for small ISPs to pay one decent admin to answer difficult questions (and indeed, my ISP is great -- the guy even helped me set up dialup PPP on a BSD box once). But the big ones just won't do that. I sincerely hope this will change. (Although I'm just fine with my small-but-excellent ISP ;-) Besides those reasons, I can think of two others, but I don't think they're serious: - Linux may not have a good reputation (it's the hacker's OS) - The ISP may want to use a proprietary dialer for some reason (Detailed accounting? Easier to setup? Some other reason?) Some thoughts: I know that the Gnome applet 'Modem Light' has a button that runs pon and poff. Is there an equivalent in KDE? I assume that other distros use ppp in the same way. Is that so? Telling them there's a stantard way to connect and solve problems with the connection would help a lot, I think. I made some comments to the proposed method, and I hope they are useful; those are the questions I'd ask if I was the manager in charge of your request. J. 1. You must have ppp installed on your computer. You must use the command line. How will the helpdesk guy know if pp is installed? How will he help the user install PPP (in all possible Linux distributions)? How will he know if the kernel supports PPP? 2. Make sure that you are logged in as root. (This is normally achieved by switching user with the su command at the command line.) Maybe there could be a problem here, but if the guy can't su-to-root, he's got bigger problems than setting up PPP... 3. Run pppconfig. Navigate the menus using the spacebar, arrow, tab, and enter keys. Will it always be available? Will it always work the same way? 13. Make sure that your modem is connected and switched on. The modem must be a serial modem and not a 'Winmodem'. How will the user know? How will the helpdesk guys know? On which serial port is it? Maybe trying to autodetect it (w/ wvdial) would help tell if it a serial modem? But -- is wvdial (or equivalent) always available? === I hope this helps somehow. I'd love to see ISPs being more Linux-friendly. J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ISP does not 'support' Linux
How will the user know? How will the helpdesk guys know? (If it's a winmodem or not) On which serial port is it? Maybe trying to autodetect it (w/ wvdial) would help tell if it a serial modem? I mean, to help tell if it's a winmodem Sorry! J. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ISP does not 'support' Linux
On Sat, Nov 30, 2002 at 12:39:25AM +1300, Richard Hector wrote: Create an XML file format for all the details required: ^^^ Yes! Buzzword! Good! :-) I don't like XML (cluttered, too verbose), but it widely accepted in the corporate world. DNS servers Dialup number Authentication type etc Write a config utility (or modify pppconfig or whatever) to read it, and only ask the remaining questions such as username and password. And being able to connect independently of which Linux distribution or which packages are installed is absolutely good. Also, if the ISP only needs to help those using this exact tool, things may become easier, but there may be other problems (see my other message). J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Calling init.d script from perl daemon?
Hi. I've spent a few hours trying and researching, but it seems that I am unable to call the init.d script for maradns (all others work). The following: my @c = (/etc/init.d/jabber, start); system @c; @c = (/etc/init.d/maradns, start); system @c; will start jabber, but not maradns. The second system call will not output anything, and will always return zero. Why does that happen? Could it have something to do with bug #164645 (maradns does not deamonize)? I was going to try a workaround (a perl script that runs as a daemon, and calls /etc/init.d/maradns when I ask it). That'd make it posible to start maradns remotely. (See the bug referreed to above) I've written the daemon, but it never starts or stops maradns. tanks, J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Easy way to modify Knoppix?
On Wed, Nov 13, 2002 at 10:43:28AM +1100, Rob Weir wrote: On Tue, Nov 12, 2002 at 11:34:12AM -0200, Jeronimo Pellegrini wrote: ll $MOUNTDIR/lib/modules/2.4.19-xfs/kernel/fs/ Stalls. And this seems to be where cp and tar stall too. Funny... I'll check this. That's quite odd. Shouldn't that directory just contain normal files? Yes. I looked again at the bugs in cloop-src, which is used to mount the compressed filesystem and found one that could explain this: the reporter says he had to reboot after trying to cp from a mounted compressed image, but this seems to be different. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Easy way to modify Knoppix?
On Tue, Nov 12, 2002 at 05:18:31PM +1100, Rob Weir wrote: Can you see what it's pausing on? Aha. ll $MOUNTDIR/lib/modules/2.4.19-xfs/kernel/fs/ Stalls. And this seems to be where cp and tar stall too. Funny... I'll check this. About the only thing I've seen cp screw up on is files in /proc, which wouldn't exist on your Knoppix disk... Maybe try the old 'tar-pipe' trick: cd destination;(cd iso_mount_point;tar -cvf - .)|tar -xf - Same thing... Blocked at $MOUNTDIR/lib/modules/2.4.19-xfs/kernel/fs/ There's also the other method (doing it from knoppix), but then I'd have to reboot every no and then... J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Easy way to modify Knoppix? [ found it ]
Ok, I think I found something (after a few days trying different keywords): http://www.es.gnome.org/cgi-bin/cvsweb/gnome-live-cd/doc/modificar_knoppix2.txt?annotate=1.1cvsroot=GNOME J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Easy way to modify Knoppix?
Hello. I'd like to put Knoppix on a rewritable CD so I'd be able to upgrade (and even change it) without wasting the CD -- but the image is too bug (700 Mb will fit on my CD-Rs, but not on my CD-RW disks). So I thought I would mount the compressed image, chroot into it, and use dpkg to change it... But since it's an iso9660 filesystem, it cannot be mounted RW. So I tried this: # Mount the Knoppix iso as downloaded, on dir work: sudo losetup /dev/loop0 `pwd`/KNOPPIX_V3.1-23-10-2002-EN.iso sudo mount /dev/loop0 work # Go to the cloop image inside work/KNOPPIX, and mount # it on /mnt/cloop sudo modprobe cloop file=`pwd`/work/KNOPPIX/KNOPPIX sudo mount -o r /dev/cloop /mnt/cloop And then I try to copy the whole tree from /mnt/cloop to some directory (./temp , for example), but the process usually goes to some point and then stalls, waiting for IO (ps shows a D flag for it). I've tried cp -a and cp -dpRPv. So... Is there an easy way to do what I'm trying to do? it'd be nice to have a Debian-based-pocket-Linux-system! :-) Thanks! J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: John the Ripper
On Sun, Oct 27, 2002 at 07:03:36PM +0200, Jerome BENOIT wrote: John writes the list of cracked passwords in ~/ which would be /root if run by the superuser. I guess that I have a basic Woody box: so why `john.pot' is in '/' but not in '/root' ? You are probably talking about John's cronjob. Since it is run by cron, $HOME/something will be /something. See bug #116257 J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: All bugs I've commented on in the Debian BTS
On Sun, Oct 27, 2002 at 03:34:52PM -0800, Mike Fedyk wrote: Hi, How do I see a list on one page all of the bugs I've commented on in the BTS? http://bugs.debian.org/from:your;email.address J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel 2.4.20 -- Where?
On Fri, Oct 04, 2002 at 11:23:19PM +0800, Russ Pitman wrote: Could someone point me to a location for debs for a 2.4.20 Kernel. 2.4.20 is not out yet -- there are only pre-releases (the latest one is 2.4.20-pre9). I believe 2.4.20 will be available in unstable a reasonably short time after it reaches its final shape. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Posted RFP, but the project seems inactive. Close bug?
Hello. As the subject says... I have posted an RPF long time ago (bug #100475), but checked the website today and it seems that there were no updates to the software since October 10, 2000 (http://sourceforge.net/projects/kuml). Should I close the bug, or just leave it there? There are better alternatives, one already packaged (argouml). J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [he's got psm] Re: can't access some https with mozilla
On Thu, Jun 27, 2002 at 06:09:10PM +0100, Ben Thompson wrote: On Thursday 27 June 2002 3:00 pm, DvB wrote: Thanks for your offer, here is the most important one (my bank) http://www.smile.co.uk If you click on account login in the top left corner, it should open a new window and a form comes up where you enter your details. Exept when I try it, all that happens is the window opens then it just times out with blank content. Same here. But when you open the source for the page, you can see it's really no big deal. And there's a link there. Did you try accessing it directly? https://welcome5.smile.co.uk/servlet/SmileBanking I tried, and I can see a form here... But I can't test if it works, because opening an account in the UK would be difficult for me right now... ;-) J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Grub Splash Image
On Thu, Jun 27, 2002 at 09:23:30PM -0400, David P James wrote: But the other day I was helping my brother with configuring grub for a multiboot involving Windows on his RedHat 7.3 box and he had a nice splash image load with grub. Is this possible with the grub that comes with Debian? Not yet... The grub homepage/manual says nothing about this option You should read the documentation that comes with the package, including the Debain-specific /usr/share/doc/grub/README.Debian: splashimage support --- Other Distributions use of graphics is an unofficial extension to Grub that has not yet been included by the developers in the official software. The patch was first proposed on the bug-grub mail list late in 2001. This feature will not be supported by upstream until after the 1.0 version of Grub is released. For more information, see bugs at http://bugs.debian.org/grub So, it does not work with Debian yet. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Frontend to System Services
On Mon, Jun 24, 2002 at 01:42:47PM -0400, Arthur H. Johnson II wrote: rcconf Yes. But see that you may get false positives because: - Some files in /etc/init.d may be marked as conffiles, and someone may have changed the init script so it does not start; - Some init scripts read a variable in /etc/default/package_name and use it to decide if it'll start or not (spamassassin/spamd and fetchmail, for example). (Or am I wrong?) J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Strange mail behaviour
sendmail -f [EMAIL PROTECTED] -F foo root gives: From [EMAIL PROTECTED] Mon Jun 24 22:27:43 2002 Date: Mon, 24 Jun 2002 22:27:42 -0100 (GMT+1) From: [EMAIL PROTECTED] (foo) To: undisclosed-recipients:; What's that doing there? man sendmail and check the -F option: -F full_name Set the sender full name. This is used only with messages that have no From: message header. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Strange mail behaviour
man sendmail and check the -F option: -F full_name Set the sender full name. This is used only with messages that have no From: message header. Oops... That make sno sense! It says From:, not To: Sorry! Anyway, I can reproduce it here. Not sure why it happens. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
I think everyone agrees that Debians package and security update systems are better. Red Hats installation procedure is userfriendlier, but that doesn't explain why professionals use it. I can think of some reasons... - Even being professionals, they want everything to be detected and configured automatically (and if not automatically, quite quickly and easily), because time == money. They can tell some unexperienced guy to install the system without worrying too much (the only time I tried this with Debian, the guy partitioned the disk exactly as I said. Then put all mount points under / -- which had 80Mb). Of course, after it's configured and running, Debian requires little attention... But then, the same is true for Red Hat (except for the security updates, of course!) - Easy-to-use graphical configuration toole. Instead of managing all possible machines, they can just help the user who asks for help on the phone. (Yes, click here, click there, now type this number) The professional may not have a problem with using vi/emacs/awhtever editor, but the usre who will need to actually have his hands on the keyboard would probably not be able to use them. - Other similar features. Allows me to focus on my work and forget as much as posible about the system internals, etc. Unfortunately, many professionals can't afford to take into consideration if the system is well-organized internally, or if it follows some strict policy, or if it is tested for a long time... They feel ok if it's tested enough, and just wait for the security updates. - Marketing related stuff. It's a big brand, and famous for being easy to use for people who don't want to know a lot about scripts system internals. - The psychological effect of having new versions of their software every 6 months. Even professionals are affected, believe me! Also, (and this is my personal experience), the professional may not care, but there'll be a lot of pressure from the users to get new versions of things... I installed Debian on several boxes where I worked. In the beginning, all was fine (we even had a local mirror, and a repository for our own debs). After a while, some of the users (and those were developers -- really not the clueless type) wanted to move to Red Hat or Conectiva. Easier to manage (so they'd focus more on developing the applications they had to develop) and with newer software (docbook-xml, spamassassin, which needs new Perl, ssh [1], and other packages -- one guy even complained about the version of Apache in potato). I question the claim that Red Hat provides better support (average helpdesk personel couldn't have helped me like (the archives of) this list have). I think there's also some psychological thing that goes on here. People think that with the help desk, they'll get an answer within a certain time, while nobody guatrantees that they'll get an answer on a mailing list. Also, you need to be careful when posting to a mailing list.. You're talking to volunteers who may just not help at all if you forget about etiquette. On the oher hand, help desk people already know that their customers may be quite angry when they call (well, the system isn't working!) Some people will of course prefer support from a mailing list, and all characteristics of Debian. Other just don't (or can't). I can't judge the system configuration system though. Consequently, I can't think of any reason for using Red Hat other than not knowing Debian (or fellow employees not knowing Debian). Well, as I said above, my experience is the opposite... Diversity is good. :-) J. [1] Although the version of ssh v1 in Debian is secure (with lots of patches), it would require other systems that talk to our system to use protocol version 1. And we can't guarantee that the Red Hat sshv1 is secure... But they needed to use it to access our network from home (or from other networks there), where the distro wasn't Debian. So I had to backport the version from Woody -- and keep following the security announces, and compiling it again every now and then. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
On Fri, Jun 07, 2002 at 11:18:01AM -0500, Dave Sherohman wrote: Probably. More significant though, is marketing. Most of us here agree that Windows isn't the best OS around, but it's got the largest userbase because of marketing and because it's what comes preinstalled on most PCs. Marketing here means a *lot*, doesn't it? Is it much of a stretch to assume that Red Hat is the most-used Linux distro because of marketing and being the most-often-preinstalled distro on PCs that come with Linux? I'd like to say marketing shouldn't be seen as just advertisement, advocacy, and that sort of thing. Just a few comments about Windows and Red Hat (and all other companies who manage to make money today). *** See that I'm not a big fan of any of them. I'm just saying that this is what they do to gain market share -- and it seems to work quite well. *** Windows may be not behave decently at all, but it sells as it is, and it's not only marketing. I can see some of the reasons: 1 - They do invest in their product, but thy'll target the users and do whatever they want.The UI, for example, that most hackers find terrible, works well for most clueless users. Microsoft gives people what they want, and will really not care about what people do not want. They didn't have to worry about security until now, for example, because their users did not complain about that (I know lots of banks that absolutely trust Microsoft products). They get statistics from their Help Desk, see what they may have missed, ask users about their products, and then do what they want. If you do what the user wants, you make it more likely that he'll stick with your product. Nothing new here -- just standard business procedure. They've just been absolutely competent in that. 2 - Development tools. This is how they built the empire. Don't exepct all developers to be brilliant. Build tools with graphical interfaces and a lot of automated stuff. No complexity -- languages like VB are just perfect for the person who just wants to see their Hello world program working. They won't understand too much about the underlying framework. These people will gradually move to making useful programs (well, if pople use them they're useful) in VB (or ASP, or wahtever). The point is: if you target the *good* developer, you won't sell too much. And if you don't sell a lot of compilers and devel tools, you won't have a lot of applications written for your OS (who's going to write the applications? You, alone?) Ironically, Microsoft has used the power of thousands of developers all over the world to build their Empire. Hm, suddenly Open Source comes to mind. Wow. Subtle. Efficient. People usually don't see this, but it's an absolutely important point. Let the clueless develop. They'll build an empire for you. Now... See that quality is not necessarily what people want. Maybe ease-of-use is a priority to them. That does not mean Microsoft isn't investing in quality. They need this so they won't get lost in a bunch of crappy and unmaintainable code. That's why they built NT -- which runs on a microkernel written by Dave Cutler (this means a lot, trust me). The win32 system on top of it may be crappy, but the kernel's gotten absolutely better. Did yo unotice that from NT to XP, Windows got gradually more stable? My bet is: they're gradually replacing crappy code with new code. I'd guess that quite soon, stability will not be an advantage of BSD or Linux over Windows. Red Hat does something similar to #1. They are obviously better than Microsoft, but still: they'll give what their users want. They'll give them graphical config tools, for example. They may not need to give their users perfect packages... Just goo denough so they won't complain too much. And one more thing they give to their users is a periodic release. New software, new versions of software... A reason for people to upgrade -- and also to be compatible with all their friends, who did upgrade already. You see... It all has to do with quality x quantity. Want lots of people to ue your product? Well, give up quality. Not pnly because you won't be able to ship quality products quick, but also because your users may show you than they're not interested in your concept of quality. What they call quality is something that workd (most of the time) and is easy to use, and has new things every now and they, so I can satisfy that little impulse inside me that makes me look for new things. I do believe, however, that it is possible to have something that does have some quality, and that lots of people will use. I didn't see MacOS/X, but it seems to be (although proprietary) something to look at. Get something that has quality and make it interesting for people. (Remember Corel, Stormix, etc?) Just thought I'd share this with you guys. :-) J. -- -- To UNSUBSCRIBE, email to
Re: Why does Mozilla modify /etc/alternatives/netscape?
Let me say something here... Many sites? Interesting. I have only found one site lately that does not work with galeon / mozilla, and it uses flash heavily. Could you elaborate on what you mean by do not work? Examples? Two banks on which I have accounts, for example... One of them will only allow IE or Netscape 4.*, and the other just won't work with Gecko (the browser just dies after a certain applet is used). Besides the banks, there are some government agencies in my country that will only support Netscape 4 (*when* they don't require Explorer at all) First, contact the webmaster for these sites. They are HORRIBLY broken if they don't support mozilla at this stage. Hm, ever tried to contact a bank about that? I did. Bank #1 was quite nice, and they said they know about the problem, but that because of the variety of Java VM implementations, they didn't manage to solve the problem yet. And they don't know when they will. Bank #2 just doesn't care. It's a w3c compliance thing. I find that when these sites are publicized, the Linux community gets active and the sites get fixed pretty quick. I've seen sites fixed quickly, and sites taht just don't answer (even when quite some people complain). Second, try altering the useragent in prefs.js. Try claiming that you are IE or old Netscape. My Mozilla crashes anyway in the first bank (which is the one I really use) :-(. I'll try with the second. Sometimes I just remove the OS specific info which they don't need anyway and things work. If it STILL doesn't work, then indeed the site is COMPLETELY broken and the webmaster should definitely be flamed. Well, a little education helps too. I agree it's broken, but the problem is that the webmasters don't agree (because it works for most users -- who use IE). J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
I'm not advocating FreeBSD. In fact, I tried it a couple of times, ran it for a week or two and hated it for a variety of reasons. Debian is the only OS/Distribution that I ever liked (which is no surprise, of course) I just wanted to say that maybe changes to stable should be more incremental. Yes, I agree... E.g., once it's determined that KDE2 is secure and stable, why not add it? Well, because the packaging could have serious problems... And because the version to be added to stable was not tested with the other packages in stable, so we don't know if it'd work. Maybe that could be remedied if the notion of proposed-updates was changed... But the developers are already discussing a lot of changes. I'd wait until Woody is released before duggesting any new ideas... J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: good Linux modems
Note this also supports the new V92 standard. It seems like a good modem. Any external modem would be Ok with Linux, but the USR models seem to be a good safe choice. USB too? I remember having heard that USB modems were basically winmodems, and therefore wouldn't work with Linux. If you confirm that USB modems are ok I'll really conside buying one. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mozilla 1.0
According to a comment on DP.org, the security patches will be applied to the version currently in woody (0.9.9, I think) instead of uploading a whole new version. What about non-security fixes? There was a lot of bug-fixing before 1.0... J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
On Thu, Jun 06, 2002 at 12:51:04PM -0400, David Z Maze wrote: This came up on debian-devel not too long ago. Someone proposed a point release to woody that would have gcc-3.1, GNOME 2.0, new KDE, and no major changes to the distribution -- even though this would require recompiling everything with a relatively untested compiler, and presenting a relatively untested desktop environment to new users. Sure, it's good PR to have a release with ooh-new-and-shiny components, but it's less clear that it'll actually *work*, which should be the point. Not only good PR... I'd say it's also useful. Using potato, we had problems using docbook XML (we had to backport the packages from unstable), to compile an ICQ jabber transport (needs gcc 3.*), and some other problems that I do not recollect immediately. But I do understand that there are packaging problems involved, of course! I don't think that blindly adding things to stable would work, and selecting what to add may require time and attention from developers... But I do think there must exist some solution to this. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: good Linux modems
wherever you heard that should be dismissed for lack of credibility. i'm running the same modem, albeit an earlier v.90 incarnation, for a year without the slightest problem. it's possible that your source made an assumption based on the fact the usr--and it's usr, not usb--site doesn't offer linux support. whatever the case may be, the usr 56k faxmodems are solid on linux. Ah, no, I meant USB (plug the modem to your USB bus...) Peolpe are telling me that although there are some that would no work, most USB modems are OK. my problem is that since my old motherboard died, I had to buy one without an ISA slot, and can't use my old modem anymore. I'm using DSL now, but am not sure for how long I can pay for the DSL line. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: in case you missed this from ponik
To me, the best solution to this would be to customize the tagline on each outgoing message, so that it would read something like you are subscribed as [EMAIL PROTECTED], to remove send a message _from that address_ to [EMAIL PROTECTED] with the magic word. That way, the clueless would have a fighting chance at getting off the list. If they are still incapable, perhaps they will include the tagline in their quoted reply so that others can take the appropriate action. Agreed. Absolutely a good thing! Wishlist bug against lists.debian.org? I don't know how hard or easy this would be to implement, but it sounds nontrivial. I suppose there are some privacy / archival issues, such as the desire to scrub mailing list archives of email addresses to foil spambots. But since the e-mail would be set to a different string for every copy that is sent, it would make more sense if the software sent it to the archives before inserting an address. But that information is available within the list emails, albeit hidden away in the full headers. Maybe better education on how to find and examine email headers in these sort of situations is in order. I woulg agree, except that I don't believe we can educate all users that get mails forwarded to them this way. They may not even know what mail headers are at all. If you refer to educating the list subscribers, then I agree. But yet -- if we blcok the guy from posting tho the list wen he start abusing, that at least helps keping the list peaceful (and saves bandwidth). In the meantime, the listmaster (or some volunteer) can help the person with the mail headers and finally unsubscribe him/her. In any case, idiots who auto-reply to every list mail they receive until they get their way are not easily defeated by any technological solution. He he... Indeed! J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
How does FreeBSD manage to stay reasonably secure and stable, yet modern (compared to Potato)? I think it's because they don't have a zero-bugs release policy like Debian. The base system is stable. The stuff in the ports tree is not, from my experience. I once decided to install gdm on a FreeBSD box... There were *lots* of broken dependencies in the ports tree, and I had to vgrep the missing dependencies in the compile logs. :-/ Besides that, Debian is an automated configuration paradise if conpared to FreeBSD. Anyway, if you manage to get over the problems you'll have to get it working, you'll find that *after* it's installed and configured, it's a very stable and powerful system. But -- they don't think twice before adding things to the ports tree. Just because it's -STABLE, that doesn't mean there can't be new software added to it. And actually, the FreeBSD -STABLE is a CVS branch! What they do periodically is to ship snapshots of it. (And of course, the snapshots are carefully prepared). J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Desktop suitability (was Re: this post is not off-topic)
I'm not really concerned with how much geeks and developers like potato for the simple reason that they (we) are capable of dealing with the uncertainties of woody/sid and might even be willing to do the occasional `./configure ; make ; make install` to get things that our distro(s) of choice don't include. For the last 18 months I worked with a very good developer (involved with the GNOME project, very competent, writes well, etc)... Since I was the sysadmin, I installed Debian in all workstations, and he started using it. He said the organization and quality of the packages was great. He really liked it in the beginning. But some time later, we all started to need stuf that was not in potato. Of course, me and hte other sysadmin compiled a lot of packages (that we needed for develpment and documentation) in a local APT repository. That other guy just didn't like the idea. It looked so ugly to him to go compiling things under demand and keeping them locally. He insisted that we were basically not using several APT features (including quick security updates), and there were packages that we were just not able to backport (because they used new Perl stuff). He considered switching his box there to Red Hat, and I had to agree it was a good idea. He didn't feel like using anything that has a testing or unstable label (and the guy is absolutely not clueless). Anyway -- my point is, new packages do make a difference, even for develpers. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mplayer (making it deb)
On Wed, Jun 05, 2002 at 04:13:01AM +1200, James Hook wrote: The mplayer people objected to distributions including it for various reasons, and there were legal problems. These may be resolved now, but The mplayer docs I have here say that there is non-GPL code which can only be distrubited in source form only. They also talk about being different for each CPU type, hence compiling from source is the 'best' option. That has changed. The mplayer website now claims mplayer is entirely GPL (they had to drop part of the code and rewrite it). Also, they have been working on CPU type detection at runtime (not perfect, but will get better). In this case I have seen packages that require a download to work, and packages that download the source code (eg. Pine) and the user must compile, has anybody investigated into if they could supply a source package version for Debian? There used to be a pine source package, but it was dropped, IIRC. I don't remember why. Personally, I think it'd be nice if APT had a fallback when you say apt-get install (if it's a source-dist-only, get the sources, compile it and install) -- with the same command used if it was a binary package. I'll think more about htis, and if it makes sense, I'll file a bug against apt. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: in case you missed this from ponik
It's not clear what can be done about this at Debian's end other than to encourage people to post full headers whenever anything goes wrong. Blocking his posts to the list while the listmaster tries to help him could help -- if the listmaster has the time to do that, of course! That would save a lot of bandwidth (the offending posts *and* the discussion about them would at least not last too long), but this can't be easily done automatically [1]. Just my 2 centavos. J. [1] Maybe with a number of votes from *subscribed* address, some automated mechanism could do that -- but sounds clumsy, abuse-prone and still not fully automatic. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bradcast2000 for Woody?
On Tue, May 21, 2002 at 06:46:56PM -0400, stan wrote: Anyone know where i can get this? Usually, you can try http://packages.debian.org/ and if that doesn't help, http://bugs.debian.org/wnpp See http://bugs.debian.org/78209 about Broadcast2000. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dpkg sanity check?
I just tried to do a sudo ls on my sid box. I'm getting a memory fault. /usr/bin/sudo has a very new file stamp and I don't remember it being updated recently. Is there a way to verify pkg's like an integrity check? Install debsums and try debsums -s sudo J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Asus A7M266 motherboard and Debian
Would you prefer these motherboards to the more recent ones from the point of view of stability under Linux? No... I always try to get information on the spcific motherboard I'm buying. And I don't trust a strong vendor name (like Asus); I instead try to find the chipsets they used, check if they have problems, and then I'll see if the vendor has traditionally done a good job integrating the chipsets in the motherboard (it doesn't matter if you're a great engineer if you're using crappy bricks or cement). I'm thinking for example of the Asus A7V266 and the A7V333, I think the former has the KT266A chipset, and the latter the KT333 chipset, both from VIA? My impression is that the more recent Athlon motherhoards have more dodgy support in general (needing to patch kernels etc,) which I don't want to get into. Well, I didn't have to patch my kernel, so no problems there. I think yes, new chipsets may take a while until the kernel works fine with them, but that's usually a short time. Hm, now that I think of it... I did patch my kernels, as I've been doing for a long time -- the kernel has one little problem with VIA chipsets in general (the system timer counter (i8253) should be reprogrammed in certain cases otherwise it may be reset to a wrong value, causing weird things to happen: gettimeofday() would return absurd values, and you'd have roblems with X blanking all the time, squid closing connections, etc). I think only the 2.4 still have the problem (the fix was introduced some time ago, but then it was removed for some reason that I don't remember). Ok. I'm just wondering in that case why it is referred to as an AMD 761 chipset. Would it not be more accurate to call it a AMD761/VIA 686b chipset? Is it just because the north bridge is more important? Thanks for replying. I don't think it's more important. Really. See, if you get a buggy chipset, or if the kernel UDMA driver for that chipset has bugs, you have a problem... I agree that it should be called AMD761/VIA686b. J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
spamassassin's whitelist_to not working...
Hello. I'm using spamassassin (2.20-1 from sid), an just noticed a problem with the whitelist_to. I'd like to whitelist_to my root address ([EMAIL PROTECTED]) because some packages will send mail ot that address, but it doesn't seem to be working. 1) If I whitelist_to something like whitelist_to [EMAIL PROTECTED] (and other domains where I can get mail) it works fine, and I can see the tag in the mail header showing that the USER_IN_WHITELIST_TO test was successful; 2) Now, if I do: whitelist_to [EMAIL PROTECTED] the test fails, and the message will go to my spam box... (I don't know why) 3) And the funny part: if I whitelist_from [EMAIL PROTECTED], it works (USER_IN_WHITELIST is there); Here's one example of mail not whitelisted: I've forged the From header to look like root only, and sent the message to [EMAIL PROTECTED], which was in whitelist_to. This is quite similar to what happens to the mails sent by the packages. From [EMAIL PROTECTED] Fri Apr 26 09:27:35 2002 Return-Path: [EMAIL PROTECTED] Delivered-To: [EMAIL PROTECTED] Received: by socrates.dnsalias.org (Postfix, from userid 1001) id 600C9431A8; Fri, 26 Apr 2002 09:27:35 -0300 (BRT) Date: Fri, 26 Apr 2002 09:27:35 -0300 From: root root To: [EMAIL PROTECTED] Subject: a Message-ID: [EMAIL PROTECTED] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.28i X-Spam-Status: No, hits=3.3 required=4.0 tests=FROM_MALFORMED,TO_MALFORMED,SIGNATURE_DELIM version=2.20 X-Spam-Level: *** Status: RO Content-Length: 5 Lines: 2 -- I'm using spamd/spamc, and whitelisting in my .spamassassin/user_prefs file, and /etc/spamassassin/local.cf is empty. I've checked the BTS and didn't find anything that helps with that. I also tried to check the sources, but it looks like the whitelist_to and whitelist-from tests are quite similar (there's one little difference, but it did look OK to me). Can anyone confirm that this happens? Any ideas? Thanks in advance! J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Asus A7M266 motherboard and Debian
On Fri, Apr 26, 2002 at 10:56:03PM -0400, Faheem Mitha wrote: [ Question about the Assu A7M266 ] Does anyone have any comments on this? If you use this motherboard on Debian without problems, particularly with a recent 2.4 kernel, I'd like to hear about it. I plan to put Woody on it with 2.4.17. Complete configuration follows. I bought it after a friend recommended it to me. I had no problems yet (using kernel 2.4.19-pre7 right now). My problems are only with the NVidia drivers (I should get a Matrox :-/ ), but the rest is OK. (BTW, I get AGP 4x here without problems) It seems to be a nice motherboard. Now, I wouldn't say it's good just because it's Asus. This time Asus did the right thing: a good AMD north bridge (because the VIA alternative would bring a bottleneck to the PCI bus, IIRC), and a reasonably good south bridge (the VIA686b, which had problems before, but not anymore, AFAIK). But I do remember some Asus motherboards that ewre really not worth their price... (some of the all-VIA boards, but I don't remember which ones) (On a side note, I am a little confused about this chipset issue. Some reviews refer to this board having the AMD 760 chipset on the northbridge, and there are also references to a VIA chipset on the southbridge. I've no idea what this terminology means.) North bridge AMD 761, south bridge VIA 686b. These are two chips you'll see on the board. As I understand, the north bridge controls the system -- it's CPU to PCI/AGP/memory bridge. The south bridge handles IO: UDMA, integrated USB, modem, audio, bridge to ISA (but you don't have ISA) If you look at the board, the north bridge is the one close to the CPU (yours will hava a heatsink), and the south bridge is closer to the other end of the board (you'll be able to see it's a VIA VT82C686b) J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compilei o kernel 2.4.16 e agora ?
On Mon, Mar 25, 2002 at 12:36:03PM -0300, Flávio Alberto Lopes Soares wrote: Olá pessoal, Neste último fim de semana consegui compilar o kernel 2.4.16 (make deps, make bzImage, make modules, make modules_install) pro meu Debian 2.2r5 Você precisa atualizar uma série de pacotes para usar kernel 2.4 com o Potato. Ponha no seu /etc/apt/sources.list: deb http://www.fs.tum.de/~bunk/debian potato main deb-src http://www.fs.tum.de/~bunk/debian potato main (O Adrian era desenvolvedor Debian até pouco tempo atrás, e vem mantendo esses pacotes há algum tempo já.) [], J. -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
