Re: graphic df type of util for Linux?
On Fri, 02 Apr 1999 00:20:41 CST, Matt Garman wrote: Until I can afford a new hard drive, I find myself typing df often to see how much free space I have on each partition. I think it would be nice to have a little utility that displays graphs of free space per partition, and updates regularly (a graphic output of df, if you will). I'm visualizing bar graphs, here, but I suppose any type of graph would work. Something that could be swallowed in some sort of desktop module would be especially nice (FvwmButtons, Wharf, ...). So maybe I could have three little bar graphs showing the free space for three partitions I specify. Does anyone know if anything of this sort exists? I use asfsm (afterstep filesystem manager) in wharf, which gives a bar graph of all mounted filesystems in a 64x64 pixel tile, and provides a utility to mount and unmount filesystems as well as showing a numerical readout of the percentage of disk space used. Unfortunately, if root is not also running asfsm it will not update the display for non root users (it shows whatever the levels were last time root ran asfsm), and it gives a constant stream of errors on the X-win initiating console for non root users of the type: sh: /usr/tmp/statfs: Permission denied I still like it better than anything else I've seen. asfsm was dropped from afterstep either at version 1.5 or the version before that (??). I salvaged mine. That's all the information I have. -- David [EMAIL PROTECTED]
Re: Star Office 5 Potato/Glibc2.1??
On Tue, 30 Mar 1999 15:11:40 CST, wrote: Hello, I was the one who posted the original message for help with this. After reading your message, I played around with the soffice wrapper that calls soffice.bin. Here's how I got mine to work... 1) I got the libc deb from slink. 2) I manually unpacked the archive (ar -x libc???.deb; tar xvfz data.tar.gz) 3) I edited the soffice wrapper: a) In the section where it sets the LD_LIBRARY_PATH (I believe it is the *) option in the block that starts with case $sd_platform in), I added the path where my glibc2.0 is located to the BEGINNING of the LD_LIBRARY_PATH. LD_LIBRARY_PATH=/home/myuser/glibc2.0/lib:$sd_inst/lib:$LD_LIBRARY_PATH b) The last line of the soffice script, I changed as follows: exec /home/myuser/glibc2.0/lib/ld-linux.so.2 $sd_inst/bin/$sd_binary $1 $2 $3 $4 $5 $6 $7 $8 $9 It took a *WHILE* (8-15 seconds) for it to load the first time (I run a P200 w/128M RAM), but it did load, and seems to function properly. When the program first loads, it gives me a message box that says, Error opening document /home/myuser/Office50/bin/soffice.bin: Nonexistent object. Filter not found. I click the OK button, and it works. That did it, *thanks!* Very well written, too. ;-) -- David [EMAIL PROTECTED] -- David [EMAIL PROTECTED]
Re: Star Office 5 Potato/Glibc2.1??
On Wed, 31 Mar 1999 09:36:23 CST, wrote: Hello: Well, now that we've all got Star Office 5.0 running on the latest cutting-edge potato systems, can anyone print from it?? When I try, I get the following messages: [..] sh: /home/myuser/Office50/glibc2/libdl.so.2: no version information available (required by sh) sh: /home/myuser/Office50/glibc2/libc.so.6: no version information available (required by /lib/libreadline.so.2) sh: /home/myuser/Office50/glibc2/libc.so.6: no version information available (required by /lib/libncurses.so.4) sh: error in loading shared libraries: /home/myuser/Office50/glibc2/libc.so.6 : undefined symbol: _dl_global_scope_end I just tried to print, and I get the same thing. Since libc6 depends on ldso, I installed that, but I restarted staroffice and it still didn't print. I'm thinking about the version information on loadable symbols option in the kernel now, but loadable modules are not much more than a black box to me, so I'm not sure how to approach that. Anyone who might know a little more? :-/ -- David [EMAIL PROTECTED]
Re: Star Office 5 Potato/Glibc2.1??
On 27 Mar 1999 23:14:05 +0200, wrote: I hope noone minds if I expand on this thread a little. Sorry, but there's too much to quote. Summary: running up to date potato, apparently glibc2.1 replaced glibc2.0 and staroffice stopped working. Jules Bean suggested a wrapper for staroffice: 1.) Get ldso from slink 2.) mkdir /usr/local/glibc2.0 3.) dpkg-deb -x ldso_1.9.10-1.deb /usr/local/glibc2.0 The original suggestion which did not work for me was to append .real to the binary and to chmod +x a wrapper script (all on two lines). #!/bin/sh LD_LIBRARY_PATH=/usr/local/glibc2.0/lib:$LD_LIBRARY_PATH /usr/local/glibc2.0/lib/ld-linux.so.1.9.10 soffice.bin.real However the staroffice binary already has a wrapper (soffice calls soffice.bin), and it looks like it might be better to add the needed code there rather than an additional layer of encapsulation. Since soffice (wrapper) has LD_LIBRARY_PATH=$sd_inst/lib:$LD_LIBRARY_PATH export LD_LIBRARY_PATH I should be able to add my glibc2.0 path: # set the glibc2.0 path glibc2.0_path=`/usr/local/glibc2.0/lib` export glibc2.0_path LD_LIBRARY_PATH=$glibc2.0_path:$sd_inst/lib:$LD_LIBRARY_PATH export LD_LIBRARY_PATH Now all that's left is to somehow tell the staroffice binary to use the ldso from glibc2.0. Originally, soffice calls the binary: exec $sd_inst/bin/$sd_binary $1 $2 $3 $4 $5 $6 $7 $8 $9 I've tried a number of ways that don't work, but is there a good way to inform soffice.bin to use /usr/local/glibc2.0/lib/ld-linux.so.1.9.10? (I've checked for dependent packages, bugs, changelogs.) Would someone please confirm that ld-linux.so.1.9.10 is the correct ldso? My potato system is showing a ld-linux.so.2 as linking to ld-2.1.1.so, which to me infers I might not have the correct ldso package to run staroffice5 (because ld-linux.so.1.9.10 might not be glibc2.0). Thanks, -- David [EMAIL PROTECTED]
synchronize docs between os'es
I keep word/staroffice docs in several different os'es (debian is my primary), and it's getting to be a PITB to manually synchronize everything across partitions. Does anyone have a simple yet elegant way of handling this? I don't want to reinvent the wheel or make this overly complex. Thanks, -- David [EMAIL PROTECTED]
Re: synchronize docs between os'es
On Mon, 29 Mar 1999 16:52:26 EST, eric Farris wrote: David Stern wrote: I keep word/staroffice docs in several different os'es (debian is my primary), and it's getting to be a PITB to manually synchronize everything across partitions. Does anyone have a simple yet elegant way of handling this? I don't want to reinvent the wheel or make this overly complex. Back when i used Windows :) i kept my docs on a FAT32 partition and mounted it when i was in linux. no syncing since they're in the same place... every night i tar/gz'd them to keep a backup (actually, i still do that...) i should put the backup on another physical machine, tho... That will work (with some partitioning), but I was thinking something more similar to however those little Personal Data Assistants synchronize their docs with the desktop box, however that works. This solution might also offer more features .. if it exists! -- David [EMAIL PROTECTED]
Re: root disk
On Fri, 12 Feb 1999 22:58:25 EST, Pete wrote: o, I figure, I'll make a root disk that my boot disk can load as a ram disk. Now, I realize I have no clue where I might find the various thing I may need to be part of this root disk. Is there anywhere I can find such a root disk, or is there some way I can reverse engineer the rescue.bin boot disk in order to extract from it the install utility? any help would be much appreciated... Hi Pete, Welcome to Debian. What I've done on an experimental basis is to take the regular boot disk (write the image to floppy), compile a new kernel with desired options, and replace the kernel on the boot disk with the new one. To find out basically what options you need to compile into the replacement kernel, read the README on the boot disk. To see the features and ramdisk Makefile mods I used for a non-scsi boot-disk, see: http://www.debian.org/Lists-Archives/debian-user-9812/msg02015.html Let me know if you have any questions. -- David [EMAIL PROTECTED]
Re: Install to disk w/school's NT present??
On Sun, 07 Feb 1999 02:12:39 PST, Kenward wrote: Odd thing in the setup. WinNT was placed in as hda1, with a small DOS logical following (hda4? IIRC). I partitioned the rest of the disk into 5 logicals inc. swap. (Where are hda2 and 3?) 1-4 are primary, 5 and higher are logical. Linux fdisk allows you to define logical partitions without having first having defined an extended partition (a type of primary partition) to contain the logical partitions. This is dubious, but usually works. DOS fdisk enforces explicitly pre-defining extended partitions before logicals are created. If you use linux fdisk to create logicals, just remember to create an extended first. Another distinction is that DOS fdisk allows writing non-sequential partition numbers (hda1 hda4), while Linux fdisk enforces sequential numbering (hda1, hda2) when writing. So, that (hda1, hda4) you have was obviously created with DOS fdisk, because linux fdisk would've written it as (hda1, hda2). This is important because partition numbers slide down when a lower numbered one of the same type (primary or logical) is deleted when using Linux fdisk. I've heard the latest commercial Partition Magic software supports partition renumbering. Linux fdisk has another oddity: you can flag multiple partitions active. Don't ask me why. What will happen if you do this is the active partition will boot. The four primary partition limit is a limitation of PC design. ide drives allow up to 63 partitions to be created, scsi drives up to15. I wanted to control the boot with LILO instead of a boot floppy. On installation was told LILO won't work on a logical, and should it be placed on hda2? I said yes. Also yes to Linux being default. hda2 is the root ( / ) partition, presumably. This is typical. An important note is that lilo only boots from primary partitions. There is a commercial boot loader, System Commander, that overcomes this limitation, but lilo does not claim to support that: /usr/doc/lilo/Manual.txt.gz--- It _can't_ be stored at any of the following locations: [..] - boot sector of a logical partition in an extended partition.* [..] *LILO can be forced to put the boot sector on such a partition by using the -b option or the BOOT variable. However, only few programs that operate as master boot records support booting from a logical partition. - So, if you want a multi-partition linux, you might have something like this: hda1 nt hda2 linux / hda3 extended hda5 linux swap hda6 linux /var hda7 linux /home hda8 linux /usr You could also just put everything on / plus swap. On reboot, set up LILO with WinNT as first in the lilo.conf list. Works perfectly. Where is hda2??? is this the MBR? Is this an iffy proposition w.r.t. stability? MBR (master boot record) is the boot record for hda (the whole drive). One advantage of installing lilo on /dev/hda2 in the above example is that you can change which os boots by just changing the active partition, which is easy from linux or dos fdisk (even a dos boot floppy with dos fdisk.) This is a very flexible setup for situations where multi-booting is desired. /usr/doc/lilo/Manual.txt.gz is terse and gangly, but a definitive resource. -- David [EMAIL PROTECTED]
Re: Install to disk w/school's NT present??
On Sun, 07 Feb 1999 10:09:53 PST, David Stern wrote: Linux fdisk has another oddity: you can flag multiple partitions active. Don't ask me why. What will happen if you do this is the active partition will boot. Of course I meant to say: .. the *first* active partition will boot. -- David [EMAIL PROTECTED]
Re: XFree86 Support for Muliheaded boxes
On Sun, 07 Feb 1999 13:11:03 CST, Chris Frost wrote: I have a Matrox Mill II and am thinking of buying a Matrox G200 to add to my box so that I could use two monitors on the same machine. Does XFree86 support this yet? I've seen talk of it for AccelX, but not really any for XFree. Answers to Frequently Asked Questions about XFree86 http://x.physics.usyd.edu.au/FAQ/index.html#TwoCards --- Q.A15- Can I use more than one video card in the same machine with XFree86? Although it is technically possible to use multiple PCI-based SVGA cards in the same machine, none of the servers currently support this. The VGA16 and Mono servers are both capable of running both a VGA compatible card and a non-VGA compatible monochrome card in the same machine. For XFree86-4.0 we are working on true multi head support. However, if you don't want to wait and you don't have the $150 to spend on Accelerated X, you might try Metro-X Enhanced Server Set 4.3, which offers mult-head support for Matrox cards, as long as only one is a Mystique, for $39, in .tgz or .rpm formats. http://www.metrolink.com/metrox/ess.html -- David [EMAIL PROTECTED]
converting numbers
Howdy, Where's the FM that tells how to convert numbers, like 0x11A to a decimal? I think there are a few common formats of numbers and I'd like to be able to recognize them and transpose them, but I don't know where to look. Thanks, -- David [EMAIL PROTECTED]
Re: converting numbers
On Thu, 28 Jan 1999 12:17:34 CST, Andrew Ivanov wrote: Howdy, Where's the FM that tells how to convert numbers, like 0x11A to a decimal? 0x11A is in hex, and to convert it to dec is 1*16^2 + 1*16^1 +10*16^0 (A=10,B=11, C=12, D=13, E=14, F=15) I should never have asked this question before having some coffee. :-) I may be wrong, but I think octal is in x0# format, so that 0x0300 would be an octal number. To convert that to dec is just 3*8^2+0*8^1+0*8^0, and ignore the leading 0 after x, which is used to idenbtify the radix. OK, hex numbers are radix 16, octal radix 8, decimal radix 1. So, a leading 0x indicates hex, and a leading 0 traditionally indicates octal, although the latter may require contextual information to distinguish between decimal, which should not be written with a leading zero, if I read Henning Makholm correctly (thanks Henning!), and I'll disregard the mention of binary. Good. Now I can convert between hex, octal and decimal. I guess I'll have to determine when to use each based on context. Thanks! Now, where is that coffee? :-) -- David [EMAIL PROTECTED]
Re: Problem install.
On Wed, 27 Jan 1999 12:15:45 +0900, wrote: I am new to Linux and trying intall Debian (hamm) to my K-62 machine. To begin with I create a rescue disk and boot. After that, following message apeared and the machine died. md driver 0.35 MAX_MD_DEV=4,MAX_REAL=8 No key (even power button) can recover this situation, so I turned off main power. I also tried another disturibution. (Turbo Linux) So no problem occured and I successfully install. In case, Same message is displayed. md driver 0.36.3 MAX_MD_DEV=4,MAX_REAL=8 what is wrong? or what is md driver? Please tells me any informations. MD driver is multiple devices. This is used for functions such as making two partitions perform as one large partition (RAID 0, etc.). Most people do not use the MD driver, but it is needed at install time. The reports I've read suggest hardware is being detected by the MD driver and an attempt is made to initialize the hardware, however either due to a system bios bug (I read reports that some system bios' were to blame) or an MD driver bug, I'm not sure which, hardware is improperly detected by the MD driver and the system hangs. Certainly you should remove any hardware that is not needed for installation. If you have a device that you suspect may cause a problem and you need it for install such as a scsi adapter for your hard drive and you can find the hardware address for it, you may be able to exclude this address at the LILO boot prompt using an exclude statement and initialize that device manually, but most reports I've seen have no such hardware. I have seen a report that the MD driver included with newer kernels may provide a solution (workaround or bugfix, I don't know). The report said that kernels 2.0.33 and 2.0.34 (in hamm) do not work, and kernel 2.0.36 (in a redhat boot disk) does work. The good news is that the MD driver was updated at kernel 2.0.35, and I'm told the slink boot disks have kernel 2.0.35, so slink boot disks (in frozen) may provide a solution. Since slink is due to release shortly, this would save you from upgrading and provide the latest software. You may want to try some other boot disks: http://www.debian.org/~adric/aic7xxx/ However, I think these are intended for improperly detected scsi adapters which hang the system during the scsi initialization phase, not the MD driver initialization phase, so they may not help. David Stern [EMAIL PROTECTED]
Re: RAID (Re: Problem install)
On Wed, 27 Jan 1999 12:05:55 +0100, wrote: David Stern dixit: MD driver is multiple devices. This is used for functions such as making two partitions perform as one large partition (RAID 0, etc.). Most people do not use the MD driver, but it is needed at install time. I recall installing the RAID option when installing Debian2.0, as many other options I did not know of at the time (such as linear, and support for filesystems I don't need). Then I had just one HDD, but now I have two. I wouldn't like the machine to try to activate RAID so, how do I get rid of it and also the rest (which I guess they'll be resource consuming)? RAID 0 utilizes multiple hard drives in tandem to improve performance. Consider /dev/hda and /dev/hdb each have a maximum 5 MB/s throughput. When utilizing both /dev/hda and /dev/hdb in tandem, 10 MB/s throughput can be achieved. However, there are drawbacks (if one drive dies, data recovery is impossible) and alternatives (partitions may be located to achieve similar results). Linear append combines partitions end to end. This is useful if a single partition of sufficient size is unavailable, or if a hard drive is very large and dividing it into smaller partitions would reduce block size and decrease wasted space. Presumably, RAID 0 on one hard drive would impose a significant performance penalty, because the hard drive must seek more times and further, and there would also be a small overhead in cpu cycles to manage the process. Linear append on one hard drive would NOT increase seeks, although there would be a small overhead in cpu cycles to manage the process, so there might be a small performance penalty. As for removing RAID or Linear Append, presumably you must copy the data to a regular partition, then update /etc/fstab to reflect the changes and reboot. Be very careful and backup first. -- David -- [EMAIL PROTECTED]
special boot disk
Hi, Early December there was a thread Debian installation hangs, and about a week later on a different thread a developer posted a url to a special disk image made just for this problem in his debian.org home directory. Now I can't find that post, or more importantly the URL. Can anyone help me out? Thanks, -- David Stern
Re: Debian goes big business?
On Fri, 22 Jan 1999 10:38:54 +0100, J.H.M. Dassen wrote: On Fri, Jan 22, 1999 at 20:26:12 +1100, Craig Sanders wrote: On Wed, Jan 20, 1999 at 06:12:14PM -0500, Ben Pfaff wrote: They should have `a word to say', and they do--they can subscribe to Debian lists and give their feedback and advice, which developers are free to follow or ignore. But they do not, and should not, IMO, have the privilege of voting or otherwise setting policy. Users are not developers and shouldn't presume to be. i mostly agree but wouldn't put it anywhere near that strongly. I would. Ben's phrasing strongly reminds me of Robert A. Heinlein; especially of the concept of TANSTAAFL and the political system he describes in Starship Troopers, where the right to vote must be earned through a tour of duty of public (not necessary military) service. In the case of Debian, users do not have the right of vote, but can earn it by becoming developers (i.e. by maintaining packages, but also by writing documentation, maintaining the website etc.). I thought it used to be that the website maintainer had no vote and that package maintainers only had to subscribe to -devel. If still true, this is not a earn voting right by public service system (per JHMD's definition) and regular direct feedback between users and developers is not actively promoted (per Ben). What's the official word on this now? I agree with Craig that Ben's view is in the right direction, but is worded too strongly. i.e.: I'm not sure when a user presumed to be a developer. I wonder if someone has been outgrowing hats lately. David Stern
Re: power outage now linux problem
On Wed, 23 Dec 1998 20:54:10 EST, AJ wrote: ok the power in my house just went out and when linux tries to boot it goes and checks my inodes and junk and then gets to a part and says: Problem: block on freelist at 06fbbc90 isnt free over adn over and doesnt boot it lets me type but i have no clue what to do.. I think what's going on is that e2fsck is looking to repair files on a block that your ide hard drive automatically checked out as bad. Since ide hard drives do this at a very low level (firmware), there's no way for you to interact without very expensive proprietary software. OTOH, if you have a scsi drive, then your controller may have software which allows you to interact with this process. If you have a backup, this would be a good time to restore. If you don't have a backup, there's a pretty small chance that any critical programs were on the bad block, so depending on how mission-critical your box is, you may wnat to forget about it or else wait a couple weeks for slink to release and then upgrade. You also may have lost data files on a bad block, but there's nothing you can do about that if you don't have a backup. Files that were partially recovered by e2fsck will be in the lost+found directory for the relevant partition, so you'll want to look in all your lost+found directories. If you want more information, see the Ext2fs-Undeletion and Partition mini-howto's. If you're new to Linux, you may want to note that ext2 filesystems are a little more sensitive than fat filesystems, so you should expect there to be some errors when the power goes out, although not always bad blocks. You also might want to consider investing in a ups. -- David -- [EMAIL PROTECTED]
Re: Quick fix + reliable editor?
On Wed, 23 Dec 1998 13:15:37 GMT, Martin Wheeler wrote: Can anyone help me with a quick fix? For various inane reasons, I'm demonstrating the possibility of using multiple, mutually-inclusive mail-readers under Linux -- mail, mailx, elm, pine, mutt, tkmail, exmh, etc. So far, so good; except that yesterday I demonstrated exmh without first making a copy of my mailbox file (/var/spool/mail/user). So exmh has now taken 2500+ mail messages and converted them into separate files, which I somehow very quickly have to get back into a monobloc form in mailbox readable format (i.e chained together with an extra cr added between each -- and I don't know of any way to do this using cat or more/less). Anyome got a handy sed/awk/perl script to do this; or any any other quick-fix suggestion? I think formail (from procmail) breaks it up, I don't know what puts it back together. Maybe you could use a for loop. $ ls filelist $ touch newfile #!/bin/bash for i in `cat filelist` do cat $i newfile echo cr newfile done --- I tested this and it worked, but I don't have any mailbox format to compare against. I'm sure more artful code exists. Second query (related; but not urgent): does anyone know of a _known_ reliable large-file text editor? One which will handle 10, 15, 25 Mb text files without hiccup, doing search--replace on the whole file without barfing, screwing up the file, etc.? (Something like Programmer's File Editor, but running under Linux). It must be known to be reliable -- i.e. tried and trusted in real-life experience, not just this should do the job. (I've been caught like that before! :( XEmacs has worked pretty well for me. I haven't done a lot of testing with files of that size, although I have done some. I've also used vi and even pico on files of that size, and they also worked well. I think PFE emulates Brief, and I think XEmacs (and Emacs) has Brief emulation as well. -- David -- [EMAIL PROTECTED]
512 node Debian CLOWN Cluster
Hello, Anyone got some details about the 512 node Debian CLOWN cluster in English? I've seen it in the Debian news section for a couple weeks, but all of the hyperlinks are in German (at least the ones I followed), and was hoping there would be an english translation or summary. I'd like to know what kind of performance a 512 node Debian CLOWN cluster achieves. (flops, mips, how quicky the equatsions were solved, ..) Also, this is pretty big cluster, so I'd be interested to know just where this sits in the record book relative to other clusters. Thanks, -- David -- [EMAIL PROTECTED]
Re: Dock an application
On Sat, 19 Dec 1998 09:25:14 GMT, Nuno Carvalho wrote: Hi, I'm using Window Maker 0.20.2-2 and I would like to know how to put applications on taskbar of my desktop! I already had lots of icons on that task but it was using dockit application and since a certain version it was abolished.=20 Any one knows how to put a simple session of rxvt on task bar, for example!? It's not so simple as drag and drop on trask bar /usr/doc/wmaker/FAQ.gz How do I get an appicon for rxvt so I can dock it? -- The default rxvt that comes with most distribtions is an outdated version of rxvt. The newest development version of rxvt is availible from ftp://ftp.math.fu-berlin.de/pub/rxvt/devel/. As of the time of this writing, the version is 2.4.6 and it natively produces an appicon without a patch. John Eikenberry has also created an rpm which is available from ftp://ftp.coe.uga.edu/users/jae/windowmaker/ As of version 2.4.6, rxvt includes WindowMaker support as well as John Eikenberry's NeXT scrollbar hack. -- David -- [EMAIL PROTECTED]
Re: Install on Adaptec 7890?
On Wed, 16 Dec 1998 05:56:01 +0200, Tapio Lehtonen wrote: Problem: Dell Precision 410 has two SCSI adapters on the motherboard, Adaptec 7880 with CD-ROM connected and Adaptec 7890 with the hard drive. Now installing Debian 2.0 fails. It sees the CD-ROM but not the hard drive. It seems this is because the Debian 2.0 Rescue disk has kernel 2.0.34, and this kernel does not support Adaptec 7890. I looked at 2.0.36 kernel, which seems to have this support. I compiled a new kernel with 2.0.36 sources, and got boot-floppies Debian package, but have not figured out how to get this new kernel on a Rescue Disk. Does some kind soul already have a Rescue disk with support for Adapted 7890? Or can someone give me some advice on making the disk. I don't know about the boot-floppies route. I read the docs, but that seemed more geared towards a distirbution release than an individual disk. However, I did successfully replace the kernel on a bootdisk yesterday, so maybe my experience can benefit you. It was actually quite simple. What I did was to select the kernel options good ol' Bruce listed in the readme on the rescue disk statically (initrd, ramdisk, loop, msdos, fat, minix, elf, ext2fs, procfs). Then I added major categories of features statically (scsi), with individual options in those categories (ai7xxx) as modules. I also used cpu type 386 to reduce the size of the kernel somewhat, although that is probably overkill. Categories which I didn't need, like isdn or ethernet, I excluded to save space, again probably overkill. Then I modified the Makefile as follows before compiling: ROOT_DEV = /dev/ramdisk RAMDISK = -DRAMDISK=1440 Then I did a make dep, make clean, and finally a make bzImage. Finally, I copied /usr/src/kernel-source-2.0.36/arch/i386/boot/bzImage to /floppy/vmlinuz (assuming you have the rescue disk mounted on /floppy). That's it. If you have any problems let me know and I'll make you a boot disk image. -- David -- [EMAIL PROTECTED]
Re: Install on Adaptec 7890?
On Wed, 16 Dec 1998 09:39:33 CST, Nathan E Norman wrote: On Tue, 15 Dec 1998, David Stern wrote: [ snip ] : Then I modified the Makefile as follows before compiling: : : ROOT_DEV = /dev/ramdisk : : RAMDISK = -DRAMDISK=1440 There's a file on the rescue disk (rdev.sh?) that contains the rdev commands you need to run on your new kernel image once you've finished compiling. No need to edit the Makefile. I find rdev annoying because it doesn't provide much in the way of feedback. Pick your poison. -- David -- [EMAIL PROTECTED]
Re: Debian installation hangs
Hi Mark, I didn't feel right about putting untested code up for the innocent to trip over, particularly newcomers, so I decided to try booting the bootdisk I created by recompiling with scsi enabled (I only have scsi drives). I experienced some errors with the ramdisk, as I feared might occur (the docs are all over the place on this issue..). So I make one modification to the Makefile regarding the ramdisk, and tested the new version, and it booted. In fact, the new version will even bootup on my scsi system, although installing without a hard drive is a dead end. I've pulled down the ill-fated first version and have replaced it with: http://weber.u.washington.edu/~dstern/resc1440_b2.bin Instructions 1.) Use the regular dd command for bootdisks: dd if=resc1440_b2.bin of=/dev/fd0 bs=512 conv=sync ; sync Substitute for the hamm bootdisk and you should at least get past the md driver message that was hanging you up, and hopefully get a full-fledged install. Let me know how it goes. --- Details of Mark's Basic Boot Disk (no scsi, no md,..) Version b2 kernel-source-2.0.34-4 gcc 2.7.2.3-4.8 Staticly linked kernel options (Yes) -- CONFIG_MODULES, CONFIG_NET, CONFIG_PCI, CONFIG_SYSVIPC, CONFIG_BINFMT_AOUT, CONFIG_BINFMT_ELF, CONFIG_KERNEL_ELF, CONFIG_M386, CONFIG_BLK_DEV_IDE, CONFIG_BLK_DEV_IDECD, CONFIG_BLK_DEV_CMD640, CONFIG_BLK_DEV_RZ1000, CONFIG_BLK_DEV_TRITON, CONFIG_BLK_DEV_LOOP, CONFIG_BLK_DEV_RAM, CONFIG_BLK_DEV_INITRD, CONFIG_INET, CONFIG_IP_NOSR, CONFIG_SKB_LARGE, CONFIG_NETDEVICES, CONFIG_MINIX_FS, CONFIG_EXT2_FS, CONFIG_NLS, CONFIG_FAT_FS, CONFIG_MSDOS_FS, CONFIG_PROC_FS, CONFIG_MOUSE, Dynamically linked kernel options (Modules) - CONFIG_BLK_DEV_FD, CONFIG_DUMMY, CONFIG_PPP, CONFIG_ISO9660_FS, CONFIG_VFAT_FS, CONFIG_NLS_CODEPAGE_437, CONFIG_SERIAL, CONFIG_ATIXL_BUSMOUSE, CONFIG_BUSMOUSE, CONFIG_MS_BUSMOUSE, CONFIG_PSMOUSE Modifications to Makefile - ROOT_DEV = /dev/ramdisk RAMDISK = -DRAMDISK=1440 Miscellaneous - $ ls -l /usr/src/kernel-source-2.0.34/arch/i386/boot/bzImage -rw-r--r-- 1 root root 321032 Dec 14 22:56 bzImage rm -f /floppy/linux cp bzImage /floppy mv /floppy/bzImage /floppy/linux --- -- David -- [EMAIL PROTECTED]
Re: query dialup-hostsname
On Tue, 15 Dec 1998 21:16:40 +0100, Marcus Geiger wrote: Hi, does anyone knows a possibility to query the hostname (dns) that will be assigned after dialing into a internet service provider ? Sometimes when logging into a ftp server, I will be greeted with something like 'dialup.a-number.lrz-muenchen.de'. I want to query this hostname to work with it in a perl script. If I understand you correctly, here's a resource which has that and a lot more. http://www.washington.edu/webinfo/cgitest.cgi -- David -- [EMAIL PROTECTED]
Re: Debian installation hangs
Hi Mark, I have some good news, and I have some bad news. THe good news first. I do have a boot disk for you. (details below) THe bad news is that since I'm removing scsi support (this was reported to be the problem, see http://www.debian.org/Lists-Archives/debian-user-9810/msg03176.html ) and since I only have scsi drives, there's no way for me to confirm that it works. I have tested it, and up to the point where it wants to see an ide drive, it works. What happens after that.. One area in which I'm less than certain is the ramdisk parameters (docs have different approaches). If you have any problems, please write down what the messages are and send them to me and/or the list and I'll see if I can fix it. So, I guess you can call this a beta release: http://weber.u.washington.edu/~dstern/resc1440_b1.bin Instructions 1.) Use the regular dd command for bootdisks: dd if=resc1440_b1.bin of=/dev/fd0 bs=512 conv=sync ; sync With any luck, you should be able to substitute this for the regular boot disk and you'll be up and running. No guarantees, and all that jazz. :) If I left anything out that you need (details below), also let me know. --- Details of Mark's Basic Boot Disk (no scsi, no md,..) = kernel-source-2.0.34-4 gcc 2.7.2.3-4.8 Staticly linked kernel options (Yes) -- CONFIG_MODULES, CONFIG_NET, CONFIG_PCI, CONFIG_SYSVIPC, CONFIG_BINFMT_AOUT, CONFIG_BINFMT_ELF, CONFIG_KERNEL_ELF, CONFIG_M386, CONFIG_BLK_DEV_IDE, CONFIG_BLK_DEV_IDECD, CONFIG_BLK_DEV_CMD640, CONFIG_BLK_DEV_RZ1000, CONFIG_BLK_DEV_TRITON, CONFIG_BLK_DEV_LOOP, CONFIG_BLK_DEV_RAM, CONFIG_BLK_DEV_INITRD, CONFIG_INET, CONFIG_IP_NOSR, CONFIG_SKB_LARGE, CONFIG_NETDEVICES, CONFIG_MINIX_FS, CONFIG_EXT2_FS, CONFIG_NLS, CONFIG_FAT_FS, CONFIG_MSDOS_FS, CONFIG_PROC_FS, CONFIG_MOUSE, Dynamically linked kernel options (Modules) - CONFIG_BLK_DEV_FD, CONFIG_DUMMY, CONFIG_PPP, CONFIG_ISO9660_FS, CONFIG_VFAT_FS, CONFIG_NLS_CODEPAGE_437, CONFIG_SERIAL, CONFIG_ATIXL_BUSMOUSE, CONFIG_BUSMOUSE, CONFIG_MS_BUSMOUSE, CONFIG_PSMOUSE Modifications to Makefile - ROOT_DEV = /dev/fd0 RAMDISK = -DRAMDISK=1440 Miscellaneous - $ ls -l /usr/src/kernel-source-2.0.34/arch/i386/boot/bzImage -rw-r--r-- 1 root root 321031 Dec 14 13:09 bzImage rm -f /floppy/linux cp bzImage /floppy mv /floppy/bzImage /floppy/linux --- -- David -- [EMAIL PROTECTED]
APC Smart-UPS 1000 and Debian
Hi, I'm not having any success with debian packaged software for my new APC Smart-UPS 1000 (listed as Linux compatible). I tried (individually) apcd, genpower, upsd. I keep getting the error message can't connect to ups or can't talk to UPS, even though I've correctly configured the software, including setting the device to /dev/ttyS1 and restarting. I have the black cable, which is supposed to be a smart model, although I don't know the exact number (# not on cable, bag, docs). At first this seemed like a cable or hardware issue, so I carefully examined my system bios and everything was good there (2F8, irq3,..). Also, I have an ATX mobo, so there are no internal connectors. Then I read the README's and FAQ's and such a little closer, and I discovered that the date of the last software revisions are really old (1-3 years). I looked in slink to see if the software was any newer, but it was the same version. Then I searched the mailing list archives and experimented with ideas found there. How can I get my APC Smart-UPS 1000 to work with Debian? Would it be possible to use PowerChute for SCO Unix with ibcs? -- David -- [EMAIL PROTECTED]
Re: APC Smart-UPS 1000 and Debian
On Fri, 11 Dec 1998 11:17:07 GMT, David Stern wrote: Hi, I'm not having any success with debian packaged software for my new APC Smart-UPS 1000 (listed as Linux compatible). I tried (individually) apcd, genpower, upsd. I keep getting the error message can't connect to ups or can't talk to UPS, even though I've correctly configured the software, including setting the device to /dev/ttyS1 and restarting. I have the black cable, which is supposed to be a smart model, although I don't know the exact number (# not on cable, bag, docs). Would you believe I had the cable plugged into the wrong com port ? :) Works great now. -- David -- [EMAIL PROTECTED]
Re: Debian installation hangs
On Tue, 08 Dec 1998 23:48:07 GMT, Mark Weston wrote: Hi, I've been trying my first ever Linux installation from a Debian 2.0 CD, and have run into problems which are beyond me at the moment. I'd be really grateful for some help. The installation (CD or rescue floppy) loads the Linux kernel, and during the long series of hardware detection messages, successfully detects all the IDE devices and then hangs on the line. md driver 0.35 MAX_MD_DEV=4 MAX_REAL=8 I've tried many times over with both media, and it always hangs in the same place. After this, I tried a RedHat 5.2 CD which installed successfully first time; on this one, the equivalent line was md driver 0.36.3etc So now I have a working RedHat install, but I'd actually set my heart on Debian. What hardware is this md driver supposed to be driving, and how possible is it to get a Debian boot/rescue disk with the later working (on my PC) version of it? BTW, at three days into being a Linux user, I don't yet feel up to the compile-your-own option. I searched the kernel sources and I found this is the multiple devices driver. That handles disk striping (RAID 0), RAID 5, etc. I have never heard of this problem, nor do I have any insight as to what might cause this error. I was taught to disconnect unnecessary devices when things like this happened, reconnect cables and check master/slave settings, but maybe you already did that. I do see they use different driver versions, and if 0.36.3 works, then perhaps you need 0.36.3 for your box. The nature of drivers is to evolve and support more devices, after all. I have kernel 2.0.36 (from slink) installed (on hamm) and md is version 0.36.3. I realize you're probably not up to compiling a kernel at this point, but you'll probably need to get a boot floppy image with the newer kernel, unless you want to wait a few weeks for slink to release. Note that I'm not sure if there are any serious implications regarding changes in the kernel between 2.0.34 and 2.0.36, or any other issues which might make this task more complex than I've thus far presumed. My experience with replacing kernels on bootdisks is nil, but if noone else is willing and you don't mind waiting a day or two, let me know. I'm not a complete moron, I know where the Bootdisk-HOWTO is, I have a fairly new hamm system with a few slink packages, so it should be doable. I wonder if there's a boot-time option to disable md. Anyone? -- David -- [EMAIL PROTECTED]
Re: installing linux debian on Toshiba Satellite
On Mon, 07 Dec 1998 22:44:20 GMT, wrote: Hi, I'm new to this list (1/2 hour). I subscribed because I intend (intended?) to install some Linux distribution on my Tosh Satellite 200CDS. Welcome! Don't go just yet. This info is just so motivating. I've never heard that called information before. I tried Suse 4.4.1 two years ago and it was only frustrating. I trashed it. Now Windows annoys me more than ever. Guys, tell me, is there a realistic chance to get Debian working on a Tosh 200CDS? I'm not a an adminstration masochist. ;-) Go to the debian mailing list archives search engine at: http://www.debian.org/Lists-Archives/ then type in Toshiba Satellite, select user (for debian-user), Oct-Dec 98, and click search. That's what I call information. Another valuable resource might be found at: Linux on Laptops http://www.cs.utexas.edu/users/kharker/linux-laptop/ There is a listing for a Toshiba 220CDS, so if a 200CDS is anything like a 220CDS, then you have an excellent chance. There are 43 Toshibas listed there. As was mentioned, you may need a zimage kernel disk image but if you make a debian-user list request I'm sure someone will respond. PS: And, 2.0 or wait a little for 2.1? Once slink is released, updating from hamm to slink should be very straightforward (not now though), so if your only concern is update difficulties, then get hamm now and update to slink after it releases. Besides, you can run most slink packages on hamm, anyway. There may be a number of dependencies to fulfill, but the Debian package finder tells you everything you need and presents it to you in an easy to use interface. http://www.debian.org/distrib/packages I've got a hamm system with some slink packages and it runs great. HTH, -- David -- [EMAIL PROTECTED]
Re: Jaz 1Gb on hamm dist
On Sun, 06 Dec 1998 16:58:01 EST, Brandon Mitchell wrote: Ack, looks like the standard dos/mac formatting trick. Try sdc4: Hi Brandon, I hate to interrupt a perfectly good thread, but can you give more details about whatever this trick is, and what the deal is with the 4th partition? I have an atapi zip drive, and if I hadn't read the messages very early in the bootup sequence with a zip disk in the drive, I never would've known to mount my zip disks on the 4th partition (/dev/hdd4). I read the zip HOWTO, but didn't find any mention of this. Mounting on the fourth partition works, but it seems wrong because fdisk doesn't show anything like that. Thanks, -- David -- [EMAIL PROTECTED]
Re: Jaz 1Gb on hamm dist
On Sun, 06 Dec 1998 21:02:56 EST, Brandon Mitchell wrote: There is some crazy way they format/partition the disks so they work on both macs and pc's, but it results in you needing to use the 4th partition on pc's. That's about all I know, except that Iomega isn't the only one to do this, Syquest does this too I think. I've got a zip disk too, which is why I'm familiar with this problem. I thought my partition table looked weird like that too. But I went ahead and repartitioned/reformated the disks for e2fs. I now see the unmaintained jazz mini-howto explains it: 6.4. Why does Iomega use partition number 4 ? Partition 4 is the default partition used by Macintosh. On a Mac, the first partition is reserved for boot info, the second for system info, the third is the resource fork, and the fourth is the data fork. Anyway, PCs and most other systems can deal with having to work on the 4th partition whereas the Mac can't deal with anything else. Iomega sends all their preformatted media with partition 4 used so that both PC's and Macs can read them and everyone avoids a big compatibility headache. (PCs with mac-disk-reading software usually expect the data to be on partition 4) Those Mac people.. :) Thanks! -- David -- [EMAIL PROTECTED]
Re: ml.org
On Sun, 06 Dec 1998 23:24:25 PST, George Bonser wrote: On Sun, 6 Dec 1998, Joey Hess wrote: Well ml.org only provided DNS, so anyone in such a situation will still ha a server with internet connectivity, they just may not have a good hostname for it. I'm I'm not sure if your offer makes sense.. They also provided dynamic DNS for people that might have had a system at home with dynamic IP. There is some software out there that has www.somename.dyn.ml.org as the home page for the program. With the demise of the dyn-DNS, these hosts will no longer be reachable. I was simply offering a temporary home until either an alternative dynamic DNS host can be found or a static address obtained. Webworkshttp://www.ddns.org DynIP http://www.dynip.com -- David -- [EMAIL PROTECTED]
Re: masq server hardware req's
On Thu, 03 Dec 1998 04:09:47 PST, Steve Lamb wrote: On Fri, Dec 04, 1998 at 12:06:10AM +1300, Michael Beattie wrote: Sorry for the dig, but Only in America America has nothing to do with it. Only in the net-idiot land of Gatesville and Windowstown. Not exactly. Christmas email reindeer animations were a conspired by a renegade faction of Santa's elves who have ties with Microsoft headquarters hometown, Redmond, WA (U.S.) city officials. In response to the U.S. Justice Department Microsoft antitrust lawsuit investigation, Redmond city mayor said regarding the allegation of ties to Santa's elves: I don't remember .. I don't know what that means. The U.S. Justice Department now believes a splinter group of the Santa's elves faction has fled to New Zealand and that Christmas email reindeer animations will increase dramatically by December 25, particularly for Debian Linux users. There is no evidence to support the rumor that Debian developers are intentionally delaying release of slink to promote wild speculation on the debian-user list. David
Re: Howto: Printing Howtos?
On Mon, 30 Nov 1998 21:13:43 MST, Bob Nielsen wrote: Joachim, I usually do something like 'zcat XXX-HOWTO.gz | a2ps -4' (although you could pipe to lpr instead). I don't usually print HOWTO's, but something I do with similarly formatted text is to use enscript to print 2-up (two pages side by side on one page) to save trees, yet remain easily readable (less page flipping, too). This yields postscript output, so you'll need to have the gs package or equivalent setup if you don't have a postscript printer. Anyway, I run a command something like: enscript -fTimes-Roman10 -2rG -p infile.txt outfile.ps (I mostly do this on other unix brands, so man enscript) and I can view the output .ps file before printing using my favorite ps viewer, xdvi, also packaged for Debian. Saves a lot of paper. -- David -- [EMAIL PROTECTED]
Re: Kernal compiling/Search Archieve and xterm ques
rathon wrote: I have two questions: 1) Is there any docs on how to do a Kernal Compile ? Download the make-kpkg package, read its documentation and use it. Oliver is the man, but I think it's called kernel-package, and the doc is /usr/doc/kernel-package/README.gz. The debian faq-o-matic covers why you should use the debian way of compiling a kernel, among other faq's: http://www.debian.org/fom/116.html -- David -- [EMAIL PROTECTED]
Re: Adjust the pop up menus in Aferstep.
On Mon, 23 Nov 1998 09:32:46 PST, Bret Craw wrote: I have added Afterstep as a window manager. When Debian installed, in the popup menu, it gives a choice of windows managers in X. After installing Afterstep, and adding it to the xinitrc file, it is the default. I would like to add it to the menus so that I can switch between WMs when I get bored. Where do I go to alter the characteristics of the pop up menu for the WMs? Theoretically you can copy /usr/doc/menu/examples/afterstep to /etc/menus/ and as root run update-menus, but I get reliable core dumps, so I adapted a menu from another wm and put it in /etc/menus/ and as root ran update-menus. -/etc/menu/AFTERSTEP ?package(afterstep):needs=wm \ section=WindowManagers title=AfterStep \ command=/usr/X11R6/bin/afterstep --- Has anyone else noticed that Midnight Commander menu entry doesn't work in afterstep and blackbox wm menus unless there is no space between the two words in the /usr/lib/menu/mc title Midnight Commander? (I'm saying that if you take the space out or fill it up with anthing other than whitespace, the menu entry works in afterstep and blackbox, otherwise it doesn't, other wm's don't mind). I've already tried adding the -u option to the mc command and parenthesizing it, and that didn't work. -- David -- [EMAIL PROTECTED]
Re: Getting Linux to recognize Windows files.
On Mon, 23 Nov 1998 09:36:56 PST, Bret Craw wrote: How do I get Linux to recognize the Windows files, so that I can use them. I have free internet access in Win98, that I can't configure for Linux. I pull down my files to the Windows partition. I have set up Linux to recognize the MSDOS partition, so that I can see what is on the hard drive. How do I get the files over into Linux and is there a way to get it to recognize the long filenames, without concatinating them? I have tried to do it with TKdesk, but it keeps telling me that the directory is not correct, even though I am looking at the file in the window. I have tried to copy the file, extracting the file to Linux, but nothing seems to work. Is there a program that may make it easier? Also, I am extremely new to Linux, and am more comfortable in XWindows. Bret Craw Maintenance Renewal Presumably you're mounting the win partitions as type dos or msdos? Try using vfat, it works for my winNT partition, and it used to work for my old win95 partition. I'm not sure about win98. Also, your kernel will have to support these fs types, so if it doesn't work, reconfigure and recompile your kernel. For a listing of fs types man mount. The win files will be owned by root, so you won't be able to write to them as a regular user. Although I'm sure there are ways to directly edit your win files as user, I prefer to need to become root so that I don't accidentally overwrite the original win file, and instead just edit local copies as user, then become root to copy them back. You should be able to copy from dos to local partition as a regular user, then as root chown the files to user.user if you want to save changes to them as user. $ su Password: # mount -t vfat /dev/hda1 /mnt/dos # exit $ cp /mnt/dos/winfile.txt . --- Note . means to right here $ su Password: # chown dstern.dstern file.txt # exit $ edit file.txt $ su Password: # cp file.txt /mnt/dos/winnt/profiles/dstern/personal/ # exit HTH, -- David -- [EMAIL PROTECTED]
Re: slink rescue disk
How to you correctly make a slink rescue disk? I downloaded resc1743.bin. First I tried superformating a 1.743 disk and running rawrite2. No go. Had trouble with anything past cyl 81. Then I downloaded resc1743 to a hamm box and ran dd if=resc1743.bin of=/dev/fd0u1743 bs=512 conv=sync ; sync. Still no go. Same problem with cyl 81+. man superformat The following example shows how to format a 1743K disk in drive 0 (83 cylinders times 21 sectors): superformat /dev/fd0 sect=21 cyl=83 -- David -- [EMAIL PROTECTED]
Bad Sector
Hi Debianeers, Today I moved some scsi partitions around and now fdisk, cfdisk and sfdisk complain that sector 3,672,497 cannot be read, and they won't allow access to that drive. My linux root partition is on the affected drive, but so far linux is running along quite well (and I've rebooted numerous times), but I was only half finished with partition relocations, so this configuration is kind of a frankenstein (mismatched partition sizes), but hey, it works. Verifying the disk media with the scsi controller did not yield any results at boot time, nor did reseating all the cables and cards. The good news is that I have a backup, albeit not super convenient (it's on a second hard drive in logical partitions, iow: not bootable). What should I do about a bad sector? The drive is only about 9 months old (ibm DCAS-34330-UW), and it's a 4.3 GB. Is this repairable? Would low-level formatting fix things, or should I be thinking about a new drive? Any ideas appreciated, David Stern fwiw: NT crashes with the INACCESSIBLE_BOOT_DEVICE blue screen of death. (I was just trying to see if I could run some adaptec diagnostic utilities.)
how do I get dselect working again?
Hi People, I'm running slink and when I tried to update my packages the other night using dselect, a number of packages, mostly perl and pam, had errors. Now dselect is stuck. I can't update, install, switch to ftp access method, .. It's already been couple days, and I've seen no mention of how to rekindle the dselect fire on debian-user. (pretty quiet, actually) What must I do to get dselect working again? I've been watching bug track and I see this remains as outstanding critical. (error messages below) Thanks, David [EMAIL PROTECTED] == If I try to access the ftp dselect method I get the error message: --- Can't locate Net/FTP.pm in @INC (@INC contains: /usr/lib/perl5/i386-linux/5.005 /usr/lib/perl5/5.005 /usr/local/lib/site_perl/i386-linux /usr/local/lib/site_perl .) at /usr/lib/dpkg//methods/ftp/setup line 7. BEGIN failed--compilation aborted at /usr/lib/dpkg//methods/ftp/setup line 7. query/setup script returned error exit status 2. --- If I try to run a install using dselect I get the error message: --- Updating package status cache...done Checking system integrity...dependency error Correcting dependencies...ok Sorry, but the following packages are broken - this means they have unmet dependencies: libpam0: Depends:libpam0g perl: Depends:perl-base --- If I try to run update using dselect I receive the error message: --- Updating package file cache...done Updating package status cache...done Checking system integrity...dependency error You might want to run `apt-get -f install' to correct these. Sorry, but the following packages are broken - this means they have unmet dependencies: perl-suid: Depends:perl libpam0: Depends:libpam0g perl: Depends:perl-base Merging Available information -- And if I try to run apt-get -f install (as suggested above): -- Updating package status cache...done Checking system integrity...dependency error Correcting dependencies...ok The following packages will be REMOVED: libmime-base64-perl alien apache-common dwww procmail-lib xaw-wrappers libgtk-perl libnet-perl premail weblint apache libmd5-perl lg-base fvwmconf dpkg-dev makepasswd kernel-package libhtml-parser-perl pkg-order ucspi-tcp-src dpkg-ftp dpkg-http econfigedit libpam0-altutil libwww-perl perl-suid debhelper info2www dftp sysutils xlogmaster libpam0 perl-tk perl WARNING: The following essential packages will be removed This should NOT be done unless you know exactly what you are doing! perl perl perl 0 packages upgraded, 0 newly installed, 59 to remove and 28 not upgraded. 7 packages not fully installed or removed. Need to get 0b of archives. After unpacking 46.8M will be freed. Do you want to continue? [Y/n] ---
Re: Voice Chatscript
On Fri, 25 Sep 1998 04:56:20 EDT, Daniel Martin wrote: [..snipped for brevity..] Hi, As Daniel points out (yeah, yeah! :), I was trying to send ascii characters to a touchtone expecting voice processing system. The solution, as Daniel also pointed out, was to make a long phone number using the comma as a (default 2 second) delay. Also as mentioned, the duration of the commas can be adjusted using the S8 register. Incidentally, the duration of the generated touchtones can be adjusted using the S11 register, just in case of a noisy line or the other end is lame and the touchtones need to be longer in duration for accurate recognition. The only thing Daniel mentioned which I'm a little uncertain about is sending +++ to the modem, because my modem book says that escapes to online-command mode. I have not used online-command mode with my modem interactively in real-time with Linux yet because I don't know how (I've done it in NT and dos using hyperterminal and smartcomm). Otherwise, Daniel is exactly right. Thanks again for the save. I should also point out that John was perfectly correct in telling me how to do what I originally asked. Thanks. I actually got that going, which is how I realized, shortly before I picked up Daniel's mail, that what I asked for was not what I wanted. Doh! All's well that ends .. David [EMAIL PROTECTED]
Voice Chatscript
Hello, I'd like my computer to call up one of those automated voice business information systems, then execute some transactions by entering my user account information and make some selections from the menu. I'll need to execute this task repetitively while I'm away from home. I tried writing a chatscript, but my modem hangs up, presumably because there's no handshake on the other end (I hear the automated voice on the other end after answering, then after about 10 seconds my modem hangs up). I see chat is intended for use with pppd, and I didn't see any options to disable the handshake expectation. ---chatscript--- ABORT BUSY ABORT NO DIALTONE ATDT777- (phone no.) \d\d\d\d\d\d\d\d\d\d (more than enough time to answer) \d\d\d\d\d99#\c (hear dialogue, enter user info followed by #) [..never gets past the initial dialogue, my end hangs up after 10 s..] So, I read the serial, net-3, isp, and other howto's, but I think all the methods employed also presume a handshake (must be implicit, it's never discussed). Then I looked at mgetty and minicom, but didn't see any way to disable the handshake expectation there either. No debugging info to see since there's no handshake .. Reluctantly, I tried my hand at the shell builtin echo command, but for some reason I don't seem to be able to get the phone number to dial, although my modem clicks. I realize this method is inferior because there is no locking of the device, but that's actually less of a concern than aborting on no dialtone, because I would like it to be able to stop redialing by leaving myself a voicemail which yields a stutter dialtone and aborts if I have ABORT NO DIALTONE set in my chatscript. ---commands as they may appear in a shell script file--- echo ATDT777- /dev/modem sleep 10 echo -c 99# /dev/modem [..never dials the phone number, just clicks once..] Finally, I looked in my Linux books, but nothing obvious there either. Anyone have a suggestion? I know I must be missing something.. David -- David -- [EMAIL PROTECTED]
Re: Voice Chatscript
On 24 Sep 1998 10:11:18 CDT, wrote: David Stern writes: I tried writing a chatscript, but my modem hangs up, presumably because there's no handshake on the other end (I hear the automated voice on the other end after answering, then after about 10 seconds my modem hangs up). I see chat is intended for use with pppd, and I didn't see any options to disable the handshake expectation. There is no implicit handshake. Chat works just as the man page says it does. Your problem is that your modem is giving up when it gets no carrier. Look in your modem manual for the AT command string to tell it not to do this. Oh yeah .. I forgot that after the handshake, CONNECT is sent. I think what I need is to override carrier detect, so I tried making my modem init string ATC0DT, but my modem still hangs up after about 10 s. ISP-Connectivity mini-howto says: Chat sends your_init_string to the modem, then dials isp_number. It then waits for CONNECT, then .., so I'm thinking that maybe chat still expects CONNECT, even if carrier detect is overridden at the modem level. That would cause no further action, as occurs. As much as I'd like to use chat so that I obtain the abort on no dialtone directive (to be able to cancel redialing remotely by leaving myself a voicemail which invokes a stutter dialtone and causes no dialtone to be activated), maybe I should just do this quick and dirty by not using a comm program, because they probably all expect CONNECT (at least I've seen no way to disable this in minicom or ckermit,) and just access the modem directly. What's the procedure for blindly sending commands to the modem? I've tried echo ATDT777- /dev/modem from the command line, but all I get is a modem click. I'm a little surprised that with all the options and params I can't just unselect the carrier detect for the comm program. Back in the days, I used to do this using DOS 2.0 on a 80286 with Hayes SmartComm, .. I think.heh heh! Hmmm .. I wonder if there is a voice auto-dialer program I can write a script file for, and if so, what would such an animal be called? Thanks John and other Debian people, David Stern [EMAIL PROTECTED]
Re: star office 4.0
On Wed, 02 Sep 1998 14:24:10 MDT, Jim Harsh wrote: I read that star office 4.0 was available for download for nocommercial users. Went to ftp.stardiv.com and downloaded what appears to be latest version, sp3-49, but it is in German. High scholl was a long time ago, and my German wasn't that good then. I've tried to go to caldera.com but my dns server can not find www.caldera.com or ftp.caldera.com. Is there anywhere else to get star office 4.0? I think debian.org has 3.1.? in nonfree/contrib/editors. I think that 49 designates the german language, and 01 designates the english language. Someone please correct me if I'm wrong. So you'd want: ftp://ftp.stardivision.com/pub/so4/linux/sp3/01/ StarOffice 5.0 is due out soon (press reports say September, but you know how that goes). Press report says 5.0 will support office 97 file formats, which might be important if you're exchanging docs from that suite. (4.0 chokes on office 97 file formats). It's a long download by modem, so make sure your ftp client has resume capability. -- David -- [EMAIL PROTECTED]
Re: exmh
I've installed it, and it looks realy cool, but it doesn't fetch my mail. I told it the name of the mail server at my ISP, but it never asked for a password and I don't see anywhere in all those bazillion configuration options where to specify it. Also, I've written a perl to check my maibox using pop3, but all that the configuration mentioned was smtp. Might this be a problem? This message is coming to you via good old netscape. exmh, like so many things in Linux, can be configured in many ways. I think the typical way to configure exmh is to have your sendmail replacement (smail, exim, ..) deliver your mail if you have a direct internet connection, or a mail delivery agent like fetchmail retrieve your mail if you're using dialup internet access. In either scenario, all exmh does is check your inbox. It doesn't retrieve messages. I think there are ways to get exmh to do this, but I don't think that is native to exmh (though I could be wrong). You should be aware that exmh is a front end (gui) for MH (or nmh, an MH almost drop-in replacement). So, you may have to configure MH a little bit as well, I know I did. That was a little bit tricky (man -k mh, and you'll see what I mean), but I still like exmh. You might also be interested to know that while MH can deliver directly to your isp's smarthost, if you have dialup internet access, that the typical configuration is to have MH pass that duty along to your sendmail replacement (smail, exim, ..). So, you see that configuring exmh might take some time and a little effort. Since I'm in the middle of exorcising the devil from sendmail, I'm probably not the best candidate to aid in this at the moment, however there's a lot of exmh help in the FAQ (built into the exmh menu), as well as at the exmh website, in the exmh mailling list archives (not as active as the debian-user list, much easier to peruse and search) and the exmh mailing list is very helpful. As I mentioned, email isn't a strong area for me, but if you don't mind a delayed response and you have any questions, let me know. David --
Re: Help on PNP cards, network config FAT 32 mounting.
On Tue, 25 Aug 1998 13:11:09 PDT, Gustavo Ribeiro Alves wrote: It's the first time I'm trying to install Linux and I'm in need of some help :). I have a PENTIUM II 300 Mhz whit a AGP video card (Appolo 65, uses the Ci rrus Logic 5465 chip), a USR Spotster 56 K Voice Faxmodem and a PCI NE2000 cl one network card. I have another computer in my house, a P100 whit the FPU an d whit the microsoft (WIN 9x) bugs :). ok I can't configure the X-windows to understand high resolution modes (+ than 8 00x600) whit my card. Although I don't really know, it sounds like you're trying to configure a fairly new graphics card which may not be supported in the version of x-windows your'e using. If that's the case, then I suggest you check out the XFree86 FAQ at the XFree86 website, see what version of X-windows supports your chipset, and then see if there's a Debian package of that version. If your card is supported with your x-windows version, and you still can't configure higher resolutions, you might find the XFree86-HOWTO (in /usr/doc/HOWTO/ or at your favorite Linux Documentation Project ftp mirror site) helpful. XFree86 used to have a really good FAQ that helped me out when I was configuring my Matrox Millenium before that was supported in a release x-windows version. I ended up compiling x-windows myself and it was fairly easy. http://www.xfree86.org/ http://sunsite.unc.edu/mdw/ My network consist of the PII and the P100. I'd like to use the IPX protocol and access whit the PII a CDROM drive on the P100 and a printer (HP DESKJET 820 Cxi - not Linux compatible). The problem is that I'd like to use the PII to access the internet via an ISP (I didn't configure it yet). There are too many different ways to configure dialup networking. Do you want dial-on-demand, a graphical utility, or just a script file? There are a number of useful HOWTO's which I suggest you look into if you plan on setting up the traditional pon/poff scripts. The NET-3 and ISP HOWTO's come to mind, and I'm sure there are others. If you'd rather use a graphical user interface, there are some debian packages you can try out. I hear wvdial mentioned from time to time here, so I presume that's popular. Personally, I like xisp because it shows you what's happening in real-time, but last time I checked there wasn't a debian package although I do recall seeing a mention that a debian developer would be packaging that in the future. After you make up your mind how you want to connect and try some stuff out and get stuck, post what your results are and what your isp expects, and there's probably someone here who can help you out. The modem is refusing to work (it's PNP) and when I tried to use the BIOS to config it, it changed my mouse configuration and I can't find the file that c ontains the mouse configuration in linux to change it. First off, do you have a win-modem? If so, you need a new modem. If not, then see if there are jumpers on your modem you can manually configure the interrupt with. I have a USR Sportster Voice/Fax 33.6 and that's what I did. I believe you can get jumper information from the USR website (or whoever bought USR, I believe http://www.usr.com still works) if you don't have the owner's manual. Manually setting the interrupts on your modem is probably the most desirable way to setup your modem, but there are other ways. I've heard of problems using the bios to config pnp devices (it never worked for me), and another option is to initialize your pnp devices in win9x and soft-boot into Linux (this worked for me, but I don't consider it a solution). The direct answer to your question is that the mouse is controlled by many means including the devices /dev/ttyS[0-3] (for COM[1-4]), and /dev/psaux for a ps/2 mouse, with a symlink from /dev/mouse to the appropriate device, as well as gpm (man gpm) for the mouse when not in x-windows, and XF86Config (man XF86Config) when in x-windows. Additionaly, you should be able to adjust some mouse settings using xset in xwindows (man xset). Next I'd like to mount a FAT32 drive and choose to boot from the linux or the win95. (I have a 800Mb partition on my HD unformated just waiting to have my old win95 installed, which I ziped and copied to another drive). I hope you know win95 will only boot from the first logical partition on the first hard drive. (or it used to, I haven't run it in years) There's also the issue of writing the master boot record. Anyways, lilo should have no problem doing this. There are so many multi-boot HOWTO's that I think I'll just refer you to those so you can try some examples. There is also a bunch of good info in /usr/doc/lilo/ as well as man lilo. You also might want to have another way of booting linux before you try doing this, i.e.: a working linux bootdisk, another bootable Linux partion, or whatever. There's also a lilo configuration utility named
Re: QUESTION
On Mon, 24 Aug 1998 20:19:29 EDT, Stephen J. Carpenter wrote: On Mon, Aug 24, 1998 at 04:47:30PM -0700, Dimas Franco wrote: HI :-) CAN I RUN PROGRAMS FOR WN95 ON LYNUX LIKE AUTOCAD 14 [..] It is possible to get CAD programs for Linux. [..] I don't know if all these links are current, and also please be warned that some are $hareware, some are commercial, at least one is only in the concept stage of development, and one or two are just toolkits and not full-fledged apps. The open source (free) contender to autocad will probably be FreeDesigner, however it is still pre-alpha. I think MicroStation, BRL and one or both of the V's are commercial autocad contenders for Linux. angela! Graphs, Diagrams http://www.mpi-sb.mpg.de/~pabst/angela/ BRL-CAD® modeling http://web.arl.mil:80/software/brlcad/ daVinci Graph Vis. Sys. http://www.informatik.uni-bremen.de/~davinci/ EnvDraw Prog. Diagrams http://www.XCF.Berkeley.EDU/~jmacd/envdraw.htm l ivtools frameworks for drawing ed. http://www.vectaport.com/ivtools/ FreeDesigner (OpenDesigner) http://www.fpa-engineers.com/OD/ FREEdraft2D http://pw2.netcom.com/~iamcliff/FREEdraft.html MicroStation http://www.bentley.com/products/microstation95/ PadDraw http://www.cs.unm.edu/pad++/paddraw/paddraw.html Scientific Applications on Linux http://SAL.KachinaTech.COM/ Sunsite Graphics Apps http://sunsite.unc.edu/pub/Linux/apps/graphics/! INDEX.html Tgif Drawing Tool http://bourbon.cs.umd.edu:8001/tgif/ Toolkit for Conceptual Modeling http://www.cs.vu.nl/~tcm/ VARIMETRIX http://www.vx.com/ Varkon http://www.microform.se/dl_linux.htm --
Re: sendmail bug?
On Fri, 21 Aug 1998 12:41:05 CDT, Nathan E Norman wrote: On Fri, 21 Aug 1998, David Stern wrote: : $ man aliases : : [..] : Loops can not occur, since no message will be sent to any person : more than once. : [..] : [..] : All I really want is to mail for root to go to dstern, and for mail to : real-root to go to root, not get looped back to dstern. How can I do : this? I can't think of an easy way to do this using only /etc/aliases and/or .forward files since aliases are edited in turn - that is real-root evaluates to root, but root evaluates to dstern. If it is acceptable, you could set an alias like this: root: root,dstern It is this situation the manpage addresses ... mail to root will be delivered to root's mailbox AND to dstern. (See? No loops : ) Right.. I now believe this is correct behavior.. however.. While testing this I discovered something which I first had difficulty believing because it contradicts my recollection, but I am forced to concede that my recollection is obviously in error; root has not received mail since I installed sendmail, and the username daemon receives root's mail. I know this because of the timestamp on root's inbox directory. I've removed every reference to the username daemon from the sendmail config files, reran sendmailconfig and restarted sendmail, I've aliased root to root, root, and yet daemon still receives root's mail. I've searched through all the sendmail documents, and I can't find any reference to what I think is happening, but I think root is being aliased to daemon. I've never heard of root being aliased to daemon, I've never even heard of the username daemon receiving mail. There are no errors in my logs and permissions and ownership of related files and directories have been quadruple verified. I did find a reference to another MTA which absolutely prohibits mail delivery to root, so I'm wondering if maybe this is a new undocumented sendmail feature? If not, is my MTA foobar, or what is going on? --/var/log/mailinfo- Aug 24 12:55:12 nitro sendmail[26499]: MAA26499: from=[EMAIL PROTECTED], size=220, class=0, pri=30220, nrcpts=1, msgid=[EMAIL PROTECTED], proto=ESMTP, [EMAIL PROTECTED] [127.0.0.1] Aug 24 12:55:12 nitro sendmail[26503]: MAA26499: to=|/usr/bin/procmail, ctladdr=[EMAIL PROTECTED] (0/0), delay=00:00:06, xdelay=00:00:00, mailer=prog, stat=Sent - I believe the procmail directive above is from root's .forward file, but how the message lands in /var/spool/mail/daemon remains a mystery. (nitro==mybox) --/var/spool/mail/daemon- From [EMAIL PROTECTED] Mon Aug 24 12:55:12 1998 Received: from nitro.net ([EMAIL PROTECTED] [127.0.0.1]) by nitro.net (8.9.1a/8.9.1/Debian/GNU) with ESMTP id MAA26499 for [EMAIL PROTECTED]; Mon, 24 Aug 1998 12:55:06 -0700 Message-Id: [EMAIL PROTECTED] X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: [EMAIL PROTECTED] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 24 Aug 1998 12:55:05 -0700 From: David Stern [EMAIL PROTECTED] -- See, it's to root, but daemon got it. What the .. --
Re: moving fvwm windows by thier handles
On Thu, 20 Aug 1998 20:59:33 CDT, the lone gunman wrote: As per my Debian 1.3 fvwm2 configuration, I could move windows by clicking and holding on any part of the border and title. That is, I could click and hold on the handles that surround a window, or the title bar and move my window around. I could only resize the window by explicit command or by those smallish L shaped corners on the windows. Now, the handles resize the window. For instance, if I display an image under xv that is big enough to push the title bar off the screen, I really can't move the window. I would like to go back to my old setup, but I do not know which fvwm2 command to use in the config file. Anyone know what I need? Thank you! Use the pager! Just drag your mondo-sized image on the pager whereever you like. That's what I ordinarily do when I'm editing an image larger than my desktop. Another option is to have a larger virtual desktop, although I don't much like that (this is done in your /etc/X11/XF86Config, man XF86Config). I thought resizing was the intended use of handles, which is why they're located on portions of the frame where it would make sense to resize; I've always moved frames using titlebars, as long as they're visible. OTOH, I don't use fvwm (unless the only alternative is KDE; heh, heh!) It sounds like your window manager desktop preferences have changed with your new package, perhaps options regarding what happens on the edge of your desktop; I don't recall any border or handle preferences. I've never experienced the behavior you're describing, but I recall there being some kind of a new configuration file scheme for fvwm derivatives, so perhaps you can specify your old config file (man fv..). You might also consider one of the fvwm configuration utilities, if they have that for bo, either the dotfile package(s) or I think I saw a new fvwm configuration utility the other day in slink..I know you have bo but you might want to have a look at it anyway. If you're having bug-like behavior, try re-installing; I've had a couple rare instances where upgrades caused titlebars and borders (hence handles) to disappear, or make all frames immovable. It doesn't sound like this is what you're experiencing, however. --
sendmail bug?
$ man aliases [..] Loops can not occur, since no message will be sent to any person more than once. [..] Loops cannot occur, ... So, if I make an alias for tsawyer to hfinn, and an alias for hfinn to dstern, and I send a mail to tsawyer, then who should get it? First I thought that meant hfinn should get it, but that clearly does not occur. When I read loops cannot occur, I get a sense of absolute, iow if I make an alias from tsawyer to hfinn, then mail sent to tsawyer goes to hfinn, no matter what alias I make for hfinn. Otherwise, a loop would occur if I were to alias dstern back to the tsawyer. If this isn't a bug, then the man page is either erroneous, or appeals to a form of logic that I don't understand. All I really want is to mail for root to go to dstern, and for mail to real-root to go to root, not get looped back to dstern. How can I do this? Someone in the know, please lend a clue to the uninitiated. Thanks. --
really-real-root (avoiding alias loop)
Hi, I've got sendmail (8.9.1) running pretty well now (woohoo!) and there are a couple finishing touches I'd like to ask for suggestions with. The first item is email aliases. I want email to root to go to a user, however when I send mail to real-root, I want it to go to the _really_ real-root, not the alias for root, which is what keeps happening. I've tried putting the real-root alias both before and after the root alias and I re-ran newaliases. Any ideas how can I avoid the loop? --
binding bind
Hi, I'm really happy my MTA is working well, now I just want to tweak these last couple details. I'm fairly new to bind, and I notice considreable delays while sending local mail (30s) and remote mail (60s) when I'm not connected to the 'net (my isp's nameservers aren't reachable) which I'd like to reduce to a more reasonable amount, or preferably eliminate. Further, these delays cause my X-win video to not be redrawn, and I can't click anything while this is happening, although I can change to another VT. Is there anyway to deactivate bind when I'm not connected to the 'net? My configuration, derived from bindconfig, is forward-only. My syslog shows named is running, purging it's cached RR's, et.al., and I'm not getting any errors. I see the /var/named/boot.options file, and I've read the /usr/doc/bin/manual/options.html, but alas I'm not sure what param(s) I should set. I'm really happy my MTA is working well, now I just want to tweak these last couple details. Thanks for any ideas you can lend. --
Re: debian 1.2 or 1.3
On Wed, 19 Aug 1998 11:13:43 CDT, Alan Maciel Salcedo. wrote: Anyone have the debian 1.3 base disk set or the 1.2 or anyone knows where to get it? thanx. http://www.debian.org says --- Old versions of Debian Debian 1.3 (codenamed bo) can be found at ftp://ftp.infodrom.north.de/pub/debian/dists/bo/. So, I click on the link, and then go to bo/disks-i386 and what do you know! --
Re: OFF-TOPIC (How do you guys sort your mail?)
On Mon, 17 Aug 1998 02:46:19 +0800, Richard L. Alhama wrote: How do you guys sort all this mail? This list is actually the first discussion-list that I've subscribed to. I'm at my wits end trying to sort all this incoming mail. I need to be able to put all the emails from debian-users to a folder (file?). I'm using pine as my MUA. What packages do I need to install? I only use Pine when my ISP makes me (on campus, it's their baby), and they haven't got procmail running yet on the server that allows imap/pop, so I don't sort mail in Pine, but my isp has this brief introduction which might be of some help to you. http://www.washington.edu/computing/faqs/plaintext/filter.intro Another, much better, reference can be found at: http://acsweb.ucis.dal.ca/fsg/notes/procmail.html I use fetchmail and procmail to get mail off the imap/pop server and sort it into my user's ~/Mail/*/ folders, then I read it with exmh (or mutt). Some of my procmail configuration files can be found at: http://weber.u.washington.edu/~dstern/linux.html#MF Further resources in my bookmarks: http://weber.u.washington.edu/~dstern/unix.html#E-Mail Thanks =8) Admiral Charah Tech Support, Cyberspace Laoag, ISP http://www2.csi.com.ph/~keyoz Overuse of the smiley is a mark of loserhood --The Jargon File V4.0.0
Open Source and DFSG conspiracy ?
Hi, Are the terms DFSG and Open Source completely interchangeable or not? They're based on identical wording. http://www.opensource.org/osd.html http://www.debian.org/social_contract.html#guidelines What's the relationship between Debian and Open Source? (I know they share webspace, but I mean business-wise.) http://www.us.debian.org/OpenSource/ http://www.opensource.org/ Does anyone know how it came to be that Debian is not listed as Open Source compliant? Open Source copied DFSG, so this appears to be in error. http://www.opensource.org/products.html http://www.opensource.org/history.html Another Eric conspiracy ? :) -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Open Source and DFSG conspiracy ?
On 09 Apr 1998 10:35:20 EDT, Ben Pfaff wrote: Hi, Are the terms DFSG and Open Source completely interchangeable or not? They're based on identical wording. Yes, they are, AFAIK. On the one hand that's nice because it's completely interchangeable, on the other hand two names for the same thing could easily be construed as misleading. http://www.opensource.org/osd.html http://www.debian.org/social_contract.html#guidelines What's the relationship between Debian and Open Source? (I know they share webspace, but I mean business-wise.) The DFSG is targeted to the free software community. Open Source is targeted at the business community. AFAIK, IIRC, and IMHO. To clarify, RedHat is DFSG-compliant (it's listed in Open Source products), and Debian is Open Source-compliant (Open Source definition is DFSG), but because DFSG targets non-commercial software (ironically the same software that appears on commercial Linux distros) and Open Source targets commercial software, this classification is correct but unintentional? shaking head What is the rationale for this distinction and why is it not stated explicitly? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Open Source and DFSG conspiracy ?
On 09 Apr 1998 13:15:21 EDT, Ben Pfaff wrote: Products that are DFSG compliant are Open Source compliant, and vice versa. It's two names for the same thing. Suits understand `Open Source' better than they understand `Free Software'; they have a suspicion of anything that has `Free' in the name. On the opposite side, hackers understand `Free' better than `Open' because they associate `Open' with abominations like `The Open Group' that aren't open at all. While I recognize that the Open Group is a commercial venture, I think the issue of `Free Software' is a much greater burden than the `Open' moniker, and any typical week's subscription to debian-devel should dispell any thoughts otherwise. Maybe hackers don't understand `Free Software' so well after all: http://www.opensource.org/history.html: [Re: `Open Source' (vs. `Free Software')] By mid-February the terminological switch was gaining momentum in the hacker community, with feedback running about 70% for it, spirited debate, and a level of awareness about the underlying issues that we found very gratifying. I think DFSG sounds crufty and legalistic while Open Source sounds clear and concise. Further, Open Source seems to be garnering recognition in the media and it seems wrong for Debian not to receive any of that recognition when Open Source is directly founded on DFSG. Therefore I think Debian deserves to both have a clear, concise software policy name (i.e. replace DFSG with Open Source) well as benefit from the recognition of Open Source. OTOH I have no experience with Debian politics, there is a freeze in effect, a constitution is being hammered out, and there seems to be no momentum here for such a proposal, so I should probably stop wasting bandwidth. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
.ps or .pdf editing app
Hi, I've tried a lot of editing apps (and conversion utilities) in hamm, but I can't find one that edits postscript or acrobat files. Does anyone know what package I can use to edit a ps or pdf form that uses Times-Roman fonts in a variety of point sizes, with some lines (no graphics)? I've tried converting to .eps and editing in tgif, however that renders the whole form illegible. I've always seen ps and pdf viewers, never editors. That just doesn't seem right. If this is off-topic, I apologize. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
randomly losing focus
Beginning about three weeks ago I've been randomly losing focus in X-windows. The way this usually occurs is that I'm navigating a menu and suddenly the (netscape, fr, exmh ..) menu disappears (usually the app will remain in focus, but not always). Further, that menu will not re-open until I open another menu first (nor will some other menus closeby). A similar type of incident occurs occasionally when I'm viewing an xterm or an auxiliary window to an app (help or config) and it suddenly loses focus, sending it underneath another app. The frequency varies, but it can be quite often, other times seldom, and it doesn't seem to depend on any particular app or number of apps. There are no bug reports for this sort of behavior in xserver-svga, and I've already verified that this is not specific to my window manager afterstep (it happens with wmaker as well, and there aren't any bugs of this sort for these two window managers). This is also not an authority issue (occurs without running su, ssl, xhost, xauth, ..). I've also looked in /var/log/messages, /var/log/syslog, however there are no errors. Losing focus while navigating menus is becoming annoying, and it no longer appears to be an unstable issue. Anyone have any ideas what might be causing this sort of behavior, or how I can go about troubleshooting it? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling kernel 2.0.33, libc6-dev
On 01 Apr 1998 12:27:13 CST, Manoj Srivastava wrote: Hi, The kernel is delibrately independent of any kernel related header files you may have installed (or that libc6 uses). It is OK to compile 2.0.33 on your machine. That's good. The newer kernel-source packages do not provide kernel-headers anymore, since the kernel-source package is architecture independent, and kernel-headers actually vary between architectures. Then does the kernel-* package information (whatever it's called) needs to be updated? --8- Description: Linux kernel source. This package provides the source code for the Linux kernel, as well as the scripts that maintain the symbolic link /usr/src/linux). This also contains everything in the package kernel-headers-2.0.33, and thus if you install kernel-source-2.0.33 you do not also need to install kernel-headers-2.0.33. --8- I still don't understand why libc6-dev depends on kernel-headers-2.0.32. I realize I'm almost certainly showing off my ignorance, but this seems highly counter-intuitive. Would someone please briefly explain how a programming language library depends on (of all things) kernel-headers-2.0.32? Call me what you will, this just seems silly. I'd like to uninstall kernel-header-2.0.32 now that I have kernel-headers-2.0.33 and kernel-source-2.0.33 installed, but dselect won't let me. Why do I still need kernel-headers-2.0.32? I smell something fishy going on here. May I recommend kernel-package package from misc? It has been designed to minimize problems during a kernel compliation. Please do read /usr/soc/kernel-package/README.gz for step by step instructions and pitfalls. I shall include the Rationale for kernel-package below I've used kernel-package several times and I think it's great. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling kernel 2.0.33, libc6-dev
On 01 Apr 1998 16:00:51 CST, Manoj Srivastava wrote: Hi, David == David Stern [EMAIL PROTECTED] writes: David Then does the kernel-* package information (whatever it's David called) needs to be updated? --8 - David Description: Linux kernel source. This package provides the David source code for the Linux kernel, as well as the scripts that David maintain the symbolic link /usr/src/linux). This also contains David everything in the package kernel-headers-2.0.33, and thus if David you install kernel-source-2.0.33 you do not also need to David install kernel-headers-2.0.33. --8 - *Groan*. Yes, I think so. Expect a new kernel-package in Incoming shortly. I didn't mean to imply that it was that important, but according to that README (see Debian's libc6 method), it looks like that should've been changed beginning with kernel-[headers,source]-2.0.32 . That seems like such a minor issue to upload all four kernel-[headers,source]-2.0.[32,33]. I think I'll set the hold in dselect. I'd read that README previously, but didn't find it consistent. Now it makes sense. Thanks. There is still one outstanding issue: dselect won't live up to the depend in libc6-dev and allow kernel-headers-2.0.32 to be uninstalled if kernel-headers-2.0.33 is installed. Why is that? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling kernel 2.0.33, libc6-dev
On Wed, 01 Apr 1998 15:11:41 PST, David Stern wrote: I didn't mean to imply that it was that important, but according to that READ ME (see Debian's libc6 method), it looks like that should've been changed b eginning with kernel-[headers,source]-2.0.32 . That seems like such a minor issue to upload all four kernel-[headers,source]-2.0.[32,33]. I think I'll s et the hold in dselect. The dreaded self-correction. Apparently, headers is alright, so I meant two packages, not four (kernel-source-2.0.[32,33]). There's also a reference to libc5-dev in the description field of both headers packages (kernel-headers-2.0.[32,33]) which looks suspicious. (Should it be libc6-dev?) -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
compiling kernel 2.0.33, libc6-dev
Hi, The 2.0.32 kernel did wonderful things for my adaptec 2940-uw, but 2.0.33 has been out for quite a while now, and I was thinking about compiling a new kernel. I'm not at all sure how libc6-dev 's dependency on stable 2.0.32 kernel-headers pertains to compiling a 2.0.33 kernel. Is 2.0.33 implicitly unstable, or can someone please clarify this matter? Also, why does the libc6-dev maintainer say that kernel-source does not provide kernel-headers, when the packages say that they do (in bug track)? Are there any other issues I should be aware of when compiling a 2.0.33 kernel, like bugs, or a new compiler I should be using? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: exim for intermittent ppp connection? Also fetchmail bombs out.
On Tue, 31 Mar 1998 23:01:58 +1000, Alan Eugene Davis wrote: I'm wondering if exim can solve my problem with mail. I am having some trouble with fetchmail. I posted to this list previously, But am uncertain that my message got through. I installed sendmail, and am uncertain I have it right. Smail worked will for quite a long time. After a recent upgrade I saw some traffic on this list about needing to reconfigure smail: that was a nightmare, having taken months when I finally did get it going. Now do it again? I'm not sure I ever understood---wrong, I'm sure I never did. Leaping from the pot to the frying pan is often not a wise approach. OTOH, it sometimes takes drastic measures to get up and running according to schedule. Right now, I don't see how replacing your MTA is going to fix fetchmail. Of course, you should be using smail from bo, unless you have a permanent internet connection..at least that's what I was told and it works for me. Sendmail is apparently working, but I'm not in command here either. Someone on the list warned that smail has been going strange places, which prompted my change: I thought it might solve my problem with fetchmail. Wrong. Fetchmail's problem --- crashing during the middle of the first message retrieval, every time, with an SMTP error because of failure to connect---still continues. I don't know how smail could fix fetchmail. Try fixing fetchmail. Since I have the latest and greatest fetchmail in hamm, and it works great (actually, I've never had a problem with fetchmail), I don't know what to tell you, other than to look at the bug reports. I looked at exim. Can exim do queues? That is, can mail be queued and send out (like runq) at the time of ppp connection? That's critical. Exim seems to be saying it's best at managing systems that are connected full time. exim is reportedly a dropin replacement for smail/sendmail, so yes (I haven't tried it, but everything I've read says this is true). I believe this is mentioned in the package description, but if not, the doc package for exim. Leaving all those aside for a moment? Is it true about smail? I've read here that smail will be fixed by the hamm release, although again I don't have any direct knowledge. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Default ObjC compiler for hamm?
On Tue, 31 Mar 1998 17:59:21 +0200, Matthias Klose wrote: Currently hamm has no working compiler for Objective-C, which can compile the GNUstep packages. GNUstep requires gcc-2.8 or egcs. On other platforms I was able to compile GNUstep with egcs-1.0.2, not so on a hamm system (with egcs (1.0.2-0.3)). The gstep-*-0.5 packages are made with gcc-2.8.1, which comes from project/experimental. Unfortunately the gcc-2.8 package is not part of hamm and is not so nice to leave gcc-2.7 installed. Well, my wish for ObjC development with hamm is to have the default compiler and a working ObjC compiler together. Would it be ok to build a gcc-objc package, which is basically gcc-2.8 without c++, which could be installed together with gcc-2.7 and conflicts with gcc-2.8? Since I recently asked off the list about this topic, I'll just post what the maintainer told me, but the bottom line is that gnustep is currently more of a technology preview than anything else, and if you just want to run the demos, you don't need gcc 2.8.1. . 8 [..my questions snipped for brevity..] Please keep in mind that this GNUstep is still in an rather early stage, and this is nothing more than a developer's release where developer means people interested in developing the GNUstep system. What has been done so far is a first implementation of Display Postscript, quite slow and slaggish, but this is been worked on. Then, the libraries that make up the OpenStep specification are worked on, i.e. the FoundationKit and the AppKit; FoundationKit is nearly complete, the AppKit is 30-50% finished (have a look at www.gnustep.org for details). This means you can write simple demo applications, i.e. with textfields, scrollers etc. pp and use DPS to draw into views. The current DGS is quite slow, not really usable yet. AFAIK, nobody has even started to implement a GUI development environment. That's certainly a very big effort. Currently, you could write applications with NeXT's tools on a machine running OPENSTEP, and you could compile them with GNUstep (only that most proably something is missing in GNUstep to compile it). So the current release just consists of a couple of libraries, a few command line utilities and daemons and a few demos that you can compile. Therefore, take these packages as a `technology preview', but don't expect to be able to actually use them for something useful. Having that said, gcc 2.8.1 is in Debian's projects/experimental directory. But you won't need it. Just install gstep-make, gstep-base, gstep-gui, gstep-xdps, dgs and gstep-xdps-examples. Then, do a source /usr/lib/GNUstep/Makefiles/GNUstep.sh, go into usr/lib/GNUstep/Apps/ and you should be able to start the examples with openapp nsbrowser.app etc. bye, Gregor 8 I ran the demos, they all ran well (but very slow). -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Design of Debian web site
On Sat, 28 Mar 1998 21:33:10 +0100, cleto wrote: On Sat, 28 Mar 1998, Thomas Apel wrote: To make it more clear what I mean: I think of home page with a navigation bar on the left and a text column on the right. The navigation bar should contain a maximum of about 6 top level categories like About, News, Support, Development... each with its second level categories. The text column could contain a brief description of what Debian is, the latest news article or other related stuff. And to make it perfect one could add a nice banner containing the logo and the text Debian GNU/Linux. All other pages could contain a smaller version of the banner and a navigation bar on the top with only the top level categories. The whole thing could be colored in white / red / dark blue like the logo. Hi, Please do not forget that there are users who access the site using text-based browsers like Lynx. Navigation bars and like make a site more difficult to navigate for them. Cleto I agree that the current Debian website design is lacking in visual impact however I also appreciate the concern for keeping web pages fast loading. I've used lynx for about five years and I don't find navigation bars prohibitive (at least per se), in fact the organization can be nice. There are some lynx-friendly web authoring guidelines at: Web authoring and Lynx http://www.crl.com/~subir/lynx/author_tips.html Rather than a single website design, why not open a Debian website design contest, with Thomas' design being the first entry? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mail and mailer questions
I recently answered a similar question: -8-- On Tue, 17 Mar 1998 18:54:35 EST, Shaleh wrote: I am using a modem and ppp to login to my ISP. What is a good setup for e-mail??? Currently I use NS4's e-mail but it is a little buggy and crashes often. I would like to have a nice X front end. I currently recieve between 250 and 350 e-mails a day and am logged in via modem sporadically at best. exmh is popular, I like it, it has never crashed on me, it has lots of nice features, and it is packaged for both bo and hamm. exmh home page http://www.beedub.com/exmh/ exmh is a tcl/tk frontend to MH or nmh (so you'll need MH or nmh too), and I'd recommend running procmail to sort your mail, and fetchmail to get your mail, and smail to send mail, assuming dialup networking. If this sounds a little more involved than configuring netscape, it is, however I think you'll find the added functionality worthwhile in the long run. I configured fetchmail and promail mostly using the documentation available in the package (man, info, /usr/doc/*), however some online procmail resources were also helpful for fine tuning: http://www.gl.umbc.edu/~ian/procmail.html http://www.ssc.com/lg/issue14/procmail.html http://www.helsinki.fi/~reriksso/procmail/mini-faq.html http://www.helsinki.fi/~reriksso/procmail/links.html The primary obstacle to overcome with smail is rewriting headers and making them stick during smtp posts. I like Daniel Martin's solution, but there are both simpler and more complex methods: http://www.math.jhu.edu/~martind/mybox.html Coincidentally, I recommend you follow the recent advice regarding MH configuration: http://www.debian.org/Lists-Archives/debian-user-9803/msg01287.html More good MH resources: http://www.cis.ohio-state.edu/hypertext/faq/usenet/mail/mh-faq/part1/f aq.html http://www.ics.uci.edu/~mh/book/ Exmh itself is pretty easy to configure, has a gui for most configuration, help menu (in 2.x version) which includes faq and mailing lists subscription. -8-- Additionally, you asked for was a text mode email client and mentioned mailbox styles, and there are at least several of each. Originally, my mail configuration was setup to work with mutt (my preferred text-mode email client) and exmh (preferred X-windows), however my mutt configuration is no longer. I believe my mutt was disavowed sometime amidst my hamm upgrade due to a long outstanding dependency. The original idea was that mutt and exmh both called on smail's post command Anyways, I just reinstalled mutt, and assigned my exmh inbox to be used, sent and received mail, and looked at my headers, and the Return-Path: field was correctly set, although the From: field apparently needs to be configured a little better because I used the example ~/.muttrc (properly modified) and html manual in /usr/doc, however sending to my isp account I get: From: [EMAIL PROTECTED] (kotsya: David Stern) If anyone knows what causes this, please post list and/or me. Anyways, Mutt has a very advanced feature list, and if you want pgp integration, you'll have to get the non-us version from nonus.debian.org (that's what someone said last night, I haven't tried it). The one thing mutt uses by default that I'm not comfortable with (yet) is vi, but this can be configured to another editor (I chose pico) in ~/.muttrc or in the EDITOR environment variable. There was also an interesting thread re: vi last night. I intend to learn vi, ultimately. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: fvwm2: icons not visible as user, ok as root.
On Tue, 24 Mar 1998 01:28:21 +1000, wrote: For quite a while, I have not been able to see most icons in fvwm2 (or other windows managers, except afterstep, which keeps icons in a certain directory). As far as I can recall, icons were visible for a brief time when I first installed Debian on this machine in July 1997. I primarily use afterstep, so I'm not a fvwm authority, but if fvwm and the other wm's use a common directory for icons, like /usr/X11R6/icons (I'm not saying it does), then you'll want to be sure your permissions are properly set for that directory. I use hamm and mine looks like: drwxr-xr-x 2 root root 5120 Feb 14 22:54 icons Did you confirm that your user(s) configuration file(s) exist with proper permissions and ownership? Did you ever modify the system config file(s) as root? If so, you may have inadvertently changed the ownership to exclude regular users. (I'm not much of an fvwm user, so I can't tell you where they're located.) Have you tried purging the package and reinstalling? Sometimes your config files will be remain, so you may have to hunt them down and manually delete them. You might try using 'updatedb' and 'locate' to help with this. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unidentified subject!
On Mon, 23 Mar 1998 16:32:41 +0100, Havas Gergo wrote: Hello I installed debian 1 months ago. I've benn trying to setup X-windows since that time. But any time I start XF86Setup after a few minutes it's frozen, and I have to login again. Sometimes it writes that some files are missing or incorrect, but does it so fast that I can't read which are those. If you have any tips or suggestions please write me. Thanks Havas Gergo If you can, switch to VC1 (virtual console #1) with CTRL-ALT-F1. You might find the messages are there. Alternatively, look in your user's home directory for the .xsession-errors file. Dotfiles (i.e. those beginning with a dot) aren't visible with regular ls, you need to use the -a param (all): ls -a. I just use filerunner (fr) and click on the show all files setting. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
NeXT Step-alike gui development tool
Hello loyal Debians, NeXT Step is a commercial un*x variant founded by Steve Jobs roughly ten years ago in his breakaway/startup company, and was purchased by Apple last year, as the basis for the upcoming Rhapsody os, which will merge with MacOS over the next couple years (yes, MacOS will be un*x). NeXT Step has lots of cool features, however one of my favorites is the gui development environment, which uses objective C with drag n drop objects. I think it is called package builder(?). Does anyone know if there is a gui development environment like this for Linux? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NeXT Step-alike gui development tool
On 21 Mar 1998 13:33:07 +0200, Tommi Kaariainen wrote: [EMAIL PROTECTED] (David Stern) writes: NeXT Step has lots of cool features, however one of my favorites is the gui development environment, which uses objective C with drag n drop objects. I think it is called package builder(?). Does anyone know if there is a gui development environment like this for Linux? There's a OpenStep clone/implementation in development, GNUstep URL:http://www.gnustep.org. Then there are Gnome URL:http://www.gnome.org/ and KDE URL:http://kde.org/. All of them have been deb-packaged. I gnu I was forgetting something. hehe. I'm downloading gnustep .5 .deb's (latest and greatest) from incoming right now. This version only released about a week ago, and already Debian has a package -- the tireless Debian developers never cease to amaze me. I know incoming .deb's don't come with a lifetime warranty, but I like to live life on the edge. (Yes, I have a backup.) I'll give kde and gnome a try after gnustep, because I hear they're not very mature yet, as you also mentioned. I didn't know any of these had a gui development environment, so thanks for all the suggestions! p.s.: Long live Debian! -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: enlightnement
On Sat, 21 Mar 1998 15:42:06 GMT, G. Kapetanios wrote: Dear All, I would like to change my window manager because I am a bit bored with fvwm2. I have chaecked enlightenment and liked what I saw. Has anayone had experience with installing it? [..] When someone asked this question four weeks ago, the reply was: -8- Date: Sun, 22 Feb 1998 03:11:48 -0500 From: Shaleh [EMAIL PROTECTED] To: debian-user@lists.debian.org debian-user@lists.debian.org Subject: Re: Enlightenment Stuff. The debs on E.org's ftp are fubarred -- avoid them. The Imlib you have IS NOT compatible w/ any E apps. Period. It has nothing to do w/ Debian or Redhat. I am the E maintainer, but my status is not complete yet. When it is expect E, imlib, and others to appear. There are a few of us working on this. To use any other E app, you MUST get imlib .11 and use it -- or wait for the other apps to get updated. All of E is in flux until the end of May, new imlib and a complete rewrite of E. Expect imlib/E to change. -8-- -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: e-mail questions
On Tue, 17 Mar 1998 18:54:35 EST, Shaleh wrote: I am using a modem and ppp to login to my ISP. What is a good setup for e-mail??? Currently I use NS4's e-mail but it is a little buggy and crashes often. I would like to have a nice X front end. I currently recieve between 250 and 350 e-mails a day and am logged in via modem sporadically at best. exmh is popular, I like it, it has never crashed on me, it has lots of nice features, and it is packaged for both bo and hamm. exmh home page http://www.beedub.com/exmh/ exmh is a tcl/tk frontend to MH or nmh (so you'll need MH or nmh too), and I'd recommend running procmail to sort your mail, and fetchmail to get your mail, and smail to send mail, assuming dialup networking. If this sounds a little more involved than configuring netscape, it is, however I think you'll find the added functionality worthwhile in the long run. I configured fetchmail and promail mostly using the documentation available in the package (man, info, /usr/doc/*), however some online procmail resources were also helpful for fine tuning: http://www.gl.umbc.edu/~ian/procmail.html http://www.ssc.com/lg/issue14/procmail.html http://www.helsinki.fi/~reriksso/procmail/mini-faq.html http://www.helsinki.fi/~reriksso/procmail/links.html The primary obstacle to overcome with smail is rewriting headers and making them stick during smtp posts. I like Daniel Martin's solution, but there are both simpler and more complex methods: http://www.math.jhu.edu/~martind/mybox.html Coincidentally, I recommend you follow the recent advice regarding MH configuration: http://www.debian.org/Lists-Archives/debian-user-9803/msg01287.html More good MH resources: http://www.cis.ohio-state.edu/hypertext/faq/usenet/mail/mh-faq/part1/f aq.html http://www.ics.uci.edu/~mh/book/ Exmh itself is pretty easy to configure, has a gui for most configuration, help menu (in 2.x version) which includes faq and mailing lists subscription. Or, you could switch to RedHat, like Bruce Parens, the former Debian maintainer, and it will probably all be configured for you. Say it isn't so, Bruce. http://www.lh.umu.se/%7ebjorn/mhonarc-files/debian-devel/msg02259.html -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lilo, linux and win95
On 18 Mar 1998 21:18:15 +1200, Carey Evans wrote: I don't think so. My PC at work, using IBM's BootManager (available in OS/2 and Partition Magic) has: - lilo.conf - boot=/dev/hda6 root=/dev/hda6 ... - The normal MBR probably can't load this though. I don't have enough experience using the MBR for the lilo boot sector to respond directly to this. I believe the lilo manual document was rewritten middle-late last year(June-September), and that there used to be some citations of unusual practices (I know there are still some good examples in there, but that's not what I'm talking about), so I don't doubt that this is possible. There was also a wish list which had more interesting ideas. Unless I've misinterpretted the lilo documentation (which coincidentally may not be up to date), your configuration seems atleast atypical, and not within the scope of documented behavior. However, you're a Linux god, and I'm a mere mortal, so for anyone reading this, consider the source! :-) I don't think it's possible to put a logical partition on a disk without it being in an extended partiton. fdisk might let you create the logical partition without explicitly creating the extended partition, but it has to do it anyway. You're inferring that fdisk implicitly creates an extended partition if a logical partition is specified, but no extended partition exists, or the logical partition cylinders are not within the boundaries of the extended partition, and that it is displayed in fdisk? That's not what I'm saying. I'm saying that not only can a logical be created without being created within an extended partition, but that it can exist without an extended partition, or at least outside of any extended partition displayed by fdisk. It's been quite a while since I've experimented with this, and I don't have a spare drive to fdisk right now, so I can't tell you how many logical partitions can be created without, or outside of, an extended partition, but even one seems inappropriate. However, since I've never seen mention of this in the fdisk documentation, I can't say for sure. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lilo, linux and win95
While most of this has been said previously, it has not been brought together and it has been less than clear at times, so I'm combining what others and myself have stated into a hopefully complete and concise summary. hda5 (or sda5) and above are logical partitions. Logical partitions cannot be used to store the lilo boot sector. For a reason I've not determined, Linux fdisk may allow logical partitions to be created and used without first creating an extended partition. I think this behavior is incorrect. My understanding is that logical partitions should be created within an extended partition, that any of the possible four primary partitions may created as extended partitions, and further that up to a possible total of 63 partitions for hda (ide) drives, and 15 partitions for sda (scsi) drives may be created. Extended partitions of the first hard disk may be used to store the lilo boot sector using the -b param, however not many programs support booting from an extended partition. For those programs that do support this option, ($ystem Commander may be one, www.v-com.com ), one or more logical partitions may effectively be booted from the extended partition lilo boot sector. /usr/doc/lilo/Manual.txt.gz - The LILO boot sector is designed to be usable as a partition boot sector. (I.e. there is room for the partition table.) Therefore, the LILO boot sector can be stored at the following locations: - boot sector of a Linux floppy disk. (/dev/fd0, ...) - MBR of the first hard disk. (/dev/hda, /dev/sda, ...) - boot sector of a primary Linux file system partition on the first hard disk. (/dev/hda1, ...) - partition boot sector of an extended partition on the first hard disk. (/dev/hda1, ...)* * Most FDISK-type programs don't believe in booting from an extended partition and refuse to activate it. LILO is accompanied by a simple program (activate) that doesn't have this restriction. Linux fdisk also supports activating extended partitions. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SOLVED: EXMH can not send e-mail (SMAIL/EXMH/MAILX)
On Sat, 14 Mar 1998 02:59:32 CST, Alexander Kushnirenko wrote: Hi, All! Thanks to everyone who replyed on this message. I think problem is fixed now, and I would like to share my little experience. [..] For Mail program (package mailx) it is done in file: /etc/smail/config visible_name=domain # !! HERE !! fake name if you wish... more_hostnames=localhost -domains hostnames=linuxbox.domain# here is the true name For EXMH/MH program one need to modify: /etc/mh/mtstailor localname: domain Thanks! This last step appears to be the the missing one I needed to make my email deliverable globally. My /var/log/smail/logfile now shows my intended hostname: Received FROM:[EMAIL PROTECTED] HOST:u.washington.edu [127.0.0.1] While an exmh FAQ might serve to get the job of configuring exmh done, and in any event would be a good addition, I wonder if a mh-config/nmh-config or exmh-config script wouldn't be even better? Actually, the smail-config could probably be improved for the internet option as well. I don't know if these ideas are in keeping with relevant traditions, or if they're the best ideas, but shouldn't configuring mail for dialup networking be fairly straightforward, no matter what an individual's preferred email client, as long as it's debianized? If there are already plans to simplify mta and email client configuration, please tell me so I don't pointlessly submit wish-list items in the bug-db. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: EXMH can not send e-mail
On Thu, 12 Mar 1998 12:35:57 CST, wrote: Hi, All I have problem sending e-mails from EXMH, but I can receive them succesfully. It looks like ordinary Mail program works fine (I can both send and receive e -mails). Here is the message from EXMH: post: problem initializing server; [RPLY] 550 '[EMAIL PROTECTED]' sen der address target 'pccmu1.fnal.gov' is not a valid e-mail domain. send: message not delivered to anyone Any suggestions what should I check? Here is a list of relevant packages: +++-===-==-== == ii smail 3.2.0.100-4Electronic mail transport system. ii exmh2.0.2-1An X user interface for MH mail. ii mh 6.8.4-17 A set of electronic mail handling programs ii mailx 8.1.1-6A simple mail user agent. If you're not running a permanent internet connection, you're gonna want to back down on the smail version, because the latest version of smail is broken for dialup networking. Try bo's smail (3.2.3, I think). That should eradicate the immediate error. If you are running a permanent internet connection, then I envy you. G. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: lilo, linux and win95
On Sat, 14 Mar 1998 02:18:13 PST, wrote: howdy all.:-) Just installed Debian linux and win95 on a system and wish to have dual boot system with LILO as boot manager. cfdisk shows: /dev/hda1 Boot Primary Dos FAT16 (has Win95 installed on it) /dev/hda5 Logical Linux /dev/hda6 Logical Linux Swap If I run LILO and tell it to generate a new /etc/lilo.conf and then answer: Install a partition boot record to boot linux from /dev/hda5?yes Install a master boot record on /dev/hadayes Make /dev/hda5 the active partition? yes I'm not intimately familiar with liloconfig (I think that's what you're using), but you can only have one lilo boot record, so take your pick -- either lilo will reside on the master boot record -OR- on the linux root partition (/dev/hda5). Whichever partition you put the boot record on, you need to set active. If you pick the mbr, say goodbye to your win95 bootloader, because lilo will overwrite the win95 bootloader (you won't be able to boot win95). Actually, there should be one backup copy to restore from. However, if you keep each os'es bootloader on their own partition, you can just change the active (bootable) partition to get the other os, should lilo fail for any reason. This, you can safely and easily do with either linux or dos fdisk, and this eliminates the problem of one boot record overwriting the other. I haven't tried it yet, but there's also a chos (choose OS) package in hamm that is supposed to make a nice menu at boot. Undoubtedly, there are numerous other possibilities, as well. The best lilo reference I found (though somewhat long and terse): /usr/doc/lilo/Manual.txt.gz There should also be a number of mini-howto's on booting possibilities: /usr/doc/HOWTO/mini/Linux+DOS+Win95+OS2.gz /usr/doc/HOWTO/mini/Linux+FreeBSD.gz /usr/doc/HOWTO/mini/Linux+NT-Loader.gz /usr/doc/HOWTO/mini/Linux+Win95.gz /usr/doc/HOWTO/mini/Loadlin+Win95.gz -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: lilo, linux and win95
Sorry but, what on earth happened to /dev/hda2, /dev/hda3 and /dev/hda4 ?? and what is the bit about Install a master boot record on /dev/hada what is /dev/h[a]da ??, shouldn't that be /dev/hda[X] ? --Jon. There doesn't need to be hda2, hda3 AND hda4. There does need to be ONE of them, because that will comprise All of the logical partitions, unless I suddently forgot everything about logical partitions. (I've done some weird things with linux fdisk that have led me to question myself on this point, but I think I'm correct.) Regarding hada, I took that as a typo.. What happened to CC:'ing the person who wrote the remarks in question? David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: lilo, linux and win95
Someone responded to me personally: David Stern, [EMAIL PROTECTED] on 3/14/98 11:44 AM If you pick the mbr, say goodbye to your win95 bootloader, because lilo will overwrite the win95 bootloader (you won't be able to boot win95). Actually, there should be one backup copy to restore from. However, if you keep each os'es bootloader on their own partition, you can just change the active (bootable) partition to get the other os, should lilo fail for any reason. This, you can safely and easily do with either linux or dos fdisk, and this eliminates the problem of one boot record overwriting the other. That's not true... You can install on the MBR and have LILO boot to Windows. I had my computer set up that way for several months before Windows ate itself and I decided I didn't want to reinstall it. Much easier than messing with partition maps every time you want to boot to another OS. This is a good point that deserves public mention. A sworn witness has testified that it's possible, and I recall it in the docs, so apparently it IS possible to let win95's boot loader and lilo's both reside on the MBR. Just don't ask me how. Everytime I tried it, windows was unbootable until I restored the MBR. After many restorations, I had chance to find the backup copy no longer valid on several occasions (requiring me to reinstall winNT), and after tiring of this, I ultimately submitted defeat. The part about running each os'es boot loader on it's own partition (not shown) still works quite well. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: CNAME records (was: Re: dynamic DNS within a non-dynamic domain)
On Mon, 09 Mar 1998 20:23:06 CST, Rich Puhek wrote: The MX record (Mail eXchange) is on your DNS server. It's what points incoming mail for username@foo.bar to the appropriate machine to handle the mail. What's worked for me is the following: mail IN A 10.0.20.1 INMX 10 mail It's probably not quite kosher (also note that these are my internal DNS numbers), but it does work. I also have an 'A' record for 'debian' (the name of our box, real imaginative) pointing to 10.0.20.1. Thanks. I get the jist of what's going on here. Further, I see Remco, who initiated this thread, has a static ip, and perhaps a semi-permanent net connection, so running his own DNS is perhaps feasible. I, OTOH, am running a dialup networking connection, and running my own DNS would not be practical (or so I've been told). This canonifying of hostnames sounded exactly like what was going on, so I read RFC-821.2 and this is entirely consistent with what seems to be happening: Once the transmission channel is established, the SMTP-sender sends a MAIL command indicating the sender of the mail. If the SMTP-receiver can accept mail it responds with an OK reply. The SMTP-sender then sends a RCPT command identifying a recipient of the mail. If the SMTP-receiver can accept mail for that recipient it responds with an OK reply; if not, it responds with a reply rejecting that recipient (but not the whole mail transaction). However, this led me to wonder how I was able to send mail with bo for so long without problems. Initially, I thought my configuration must have changed, but recently I've realized that implementation of the latest anti-spam MTA features across the 'net roughly coincided with my hamm upgrade, thus it may not be my configuration that changed -- I'm not sure. At least this narrows the problem down to C_NAME lookup at my SMTP server send-time. Now all I have to figure out is how to alleviate that. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: CNAME records (was: Re: dynamic DNS within a non-dynamic domain)
On 09 Mar 1998 10:50:15 +0100, wrote: Note that AFAIK all SMTP mailers (well, sendmail at least) will rewrite the addresses in the mail headers. That includes rewriting a CNAME pointer to the canonical name it points to. If this is what ails my smail, then the vast majority of domains have recently and unanimously conspired to reject my mail, because I've been sending mail just fine up until I installed hamm. This idea seems an incomplete explanation, if this also applies to me, but anything's possible.. So even if you send out mail with the CNAME in the From: header field, the recipient mailer will rewrite it.. no way around it. The correct way to solve this is to not use CNAMEs but to just set up an extra A + MX record. Or perhaps just an MX record if all you're using it for is mail. I don't run sendmail, but what's going on sounds an awful lot like like what I'm experiencing with smail, so I'd like to know more about setting up an MX record, whatever that is.. Where can I find out about this? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: XF86Setup
On Mon, 09 Mar 1998 19:29:44 EST, Taylen wrote: Hey...I've been trying for a few days now to get x windows to work, I tried using XF86Setup to make my configuration file, but the graphics for it were screwed up and unreadable...I then tried to use just XF86Config and created an X configuration filehowever, when I tried to start x, the screen was still messed up as it had been when I had run XF86SetupI haven't been able to figure out what the problem might be, does anyone know what might be causing this problem? Is this a new install? What video card are you running? If you're running a Matrox Millenium II, you're gonna have to wait for the .deb, because the latest XFree86 version with support just released a few days ago, or so it was reported here last night, I believe. Of course you can compile it yourself. I had to do this when my Matrox Millenium was unsupported, and it was pretty painless for me, even though I was totally new to Linux. (tar -xvzf, read README, edit Makefile, make, make install) You also might try the excellent XFree86 FAQ at: http://x.physics.usyd.e du.au/FAQ/index.html The XFree86 Matrox FAQ is at: http://matrox.alloy.net/eng/faq.html If you've got a laptop, there are special resources for you too, but right now the matrox seems to be the most FAQ. I hope you're also reading the man pages, and /usr/doc/X11R6/lib/X11/doc/README* . -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: installing debian problem
On Tue, 10 Mar 1998 23:10:10 +0800, kaiwei wrote: what is a image file? It's an image of a disk (each track is copied, even if empty). To use it, you'll need the utility rawrite.exe, which can be found in the /dosutils directory. Read the README file which accompanies it for more info. in the inital boot medium, i use loadlin to boot install from a dos system. i used the command:- loadlin linux root=dev/ram initrd=root.bin but i got a error message of:- not a image file please enter name of kernel image file followed by optional commmand line parametersf for linux the linux file i got was from the debian ftp site. (normal linux file) I have no practical experience with loadlin. also i'm using win95 also. if i install debian, can i keep it use it too? will it affect it in anyway. i have set aside a empty partition of 1 gb free for debian. is it enough? how do u create a swap partition? from the same partition? i spent 3 days working on this i still can't get it to work What I recommend is to put the lilo on the Linux / (i.e.: the root partition). By default, lilo installs onto the MasterBootRecord, where the win95 boot record lives. If you accept this option, it will write over your win95 bootloader. There may be a backup copy of the MBR you can restore (or not, depending), but if you put lilo on the linux root partition, each bootloader has it's own space. Then you can make an entry for win95 in the lilo menu, so you won't need to change the bootable partition under ordinary conditions. Don't forget to flag the lilo partition as bootable. You create all your partitions during installation using the Linux fdisk utility, which has some built-in help. Here are some docs which thoroughly cover installation: Installing Debian ftp://ftp.us.debian.org/debian/stable/disks-i386/current/install.html Debian Linux User's Guide http://www.linuxpress.com/001001.htm Debian Installation and Getting Started http://www.ssc.com/lg/issue15/debian.html -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
smail, the sequel
After reading and trying out many of the ideas I've seen here and elsewhere regarding smail configuration for dialup networking, I've decided to ask for some more help. VISIBLE_NAME and PRIMARY_NAME are now~ being set to u.washington.edu after PRIMARY_NAME=u.washington.edu properly configuring hostnames and VISIBLE_NAME=u.washington.edu visible_name in /etc/smail/config ,~ Likewise, the Sender: field in a mail to my isp account is okSender: [EMAIL PROTECTED] now, thanks: Moreover, I think I found the explanation as to why mails to my isp account have the correct hostname in the Sender: field while perusing man smailtrns, although I'm not certain. My theory is that by setting visible_name to my isp's domain, the mail is considered local and there is no lookup with the dns. Ultimately, I have the problem with most non-local mail: ~~~ /var/log/smail/logfile) ~ Received FROM:[EMAIL PROTECTED] HOST:localhost [127.0.0.1] [..] ~~ And the response from my smarthost: From: [EMAIL PROTECTED] (evil looking picon)~ |- Failed addresses follow: ---| [EMAIL PROTECTED] ... transport smtprewriter: 550 '[EMAIL PROTECTED]SIZE=724' sender address target 'localhost' is not a valid e-mail domain. |- Message text follows: --| Received: from localhost ([127.0.0.1]) by localhost with esmtp (ident kotsya using rfc1413) id m0yBxrw-0001bOC (Debian Smail-3.2 1996-Jul-4 #2); Mon, 9 Mar 1998 00:25:08 -0800 [..] ~~ Would someone with more experience in this area please advise me what I need to change to be able to use smail to send mail somewhere other than myself and this list? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: Mostly fresh start install?
On Mon, 09 Mar 1998 16:54:22 EST, T-SNAKE wrote: I want to mostly start anew with my install, since I think during my software install, I hit some wierd button accidentally and it FLEW through a bunch of prompted questions and I think totally screwed sendmail... so I'm gonna start over. BUT, I'm not sure what the best way to proceed is. Isn't there a sendmailconfig (I forget the exact name)? Use it. If you want to re-run the sendmail post-install script, you can find it in /var/lib/dpkg/info/sendmail*. Should I just re-initialize my partitions (with the rescue disk), which assuming will delete the info on them? Or do I need to do something more drastic? The only partition change I want to make is I (I don't know why) a 32MB and 16MB partition for swap, but might as well combine them for one 48MB swap partition I don't know why you did that either :-). You should be able to disable your swap partitions (man swapoff) and then fdisk delete the two swap partitions, and then create a new swap partition. The only thing I'm not sure about is how to format a swap partition. No need to re-install. The cliche I always heard regarding Linux re-installs was: this isn't win95. Don't forget to update your /etc/fstab to show the correct partition(s) before you reboot. I've re-installed after using development kernels, compiling a lot of my own software, and following upgrades on a more popular (and flaky) Linux distro, but I haven't found the need with Debian. I even upgraded to hamm without problems (the hamm mta is hosed imho, but that's another story..). Any reccomendations? I'm NOT using HAMM. I believe I am using 1.3.1 or something like that... peace, If these are truly your only reasons for starting over, I'd say don't do it, because you'll have to reconfigure your printer, X-windows, and all the rest needlessly. The cliche you'll probably hear is: Linux isn't win95. IOW: just because you had to reinstall win95 doesn't mean you need to reinstall Linux; you just need to reconfigure. This can usually be done without rebooting, by stopping and restarting services (I don't want to tell you wrong, and I don't recall if bo was different in this area.) FWIW, I even resize my / , /var , /home , /usr partitions by copying them elsewhere, resizing them with fdisk, reformatting, and then copying them back. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: fdisk and mount (was Re: Dselect)
On Sat, 07 Mar 1998 20:42:53 +0100, Hengstberger wrote: Hi! I`m bloody new with linux.Partitionig my harddrive I lost 0,5GB, thant means it's neither in the dos partitioning list nor in the Linux's. I don't really need it at present. How can I access this space again, without competely formatting the entire HDD? I don't know how you determined that a half gigabyte was lost, but you can look at your linux partition table with fdisk -l (as root, man fdisk). THis will show the beginning and ending cylinders, so if there is a discontinuous region, you can use fdisk (or cfdisk) to add a new partition (man fdisk). Note that you cannot enlarge an existing partition, although you can delete a partition and replace it with a larger one -- all data on that partition will be lost, and you will have to format new partitions (man mke2fs). Prior to using fdisk you should be aware of the destructive nature and backup anything which is invaluable. If you make modifications with fdisk, don't forget to update /etc/fstab to show the new partition numbers for the next time you reboot. If all of your cylinders are accounted for, consider that mke2fs reserves 5% for root by default. I've tried to mount my CD-ROM (ID 2, hdc ) any recommended way e.g. mount -t ISO 9660... . Can anyone tell me any alternatives that will work better? First you need a directory where the cdrom is to be mounted. mkdir /cdrom Then, if you put an entry in /etc/fstab, you only need to specify the device, (you may the filesystem type and mount location): --add to /etc/fstab-- /dev/hdc /cdromiso9660 noauto 0 0 - and then you should be able to just type: mount /dev/hdc (the long way is to type mount -t iso9660 /dev/hdc /cdrom) There are many variations on this, ranging from setting up aliases in your $HOME/.bashrc file, to making a group of priviledged users for mounting removable media. Since you're new to Linux, I also recommend you try out one of the many Linux reference books, possibly available online, at your bookstore or library. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: afterstep
On Tue, 03 Mar 1998 14:16:41 PST, Luiz Otavio L. Zorzella wrote: I upgraded afterstep to 1.4, but I had to downgrade it back to 1.0, cuz the new one did not get any of my old confs, and looked horrible! Did I do anything wrong? You need to follow the instructions in: /usr/doc/afterstep/README.gz I didn't try it, but there's a -f param if you want to use your crufty old .steprc , not that you would -- the new file layout is exceptional. You can skip that part about the security hole if you give each user their own copy of the GNUstep/library/afterstep (don't cp startmenu to /usr/lib/afterstep). If sloppy focus is new and that bothers you, you can change that in: ~/GNUstep/Library/AfterStep/desktop/feels/feel.Debian (by uncommenting # ClickToFocus, commenting SloppyFocus) You may also want to modify: (assuming 16bpp, adjust accordingly) ~/GNUstep/Library/AfterStep/base.16bpp (for icon locations), ~GNUstep/Library/AfterStep/autoexec.16bpp (for startup apps), and ~/GNUstep/Library/AfterStep/wharf (for .. you guessed it) Well, back to smail fun.. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: 80 MB ram in debian system
On Mon, 02 Mar 1998 10:18:16 CST, Russ Cook wrote: Hi All, This weekend, I added some ram to my system. I now have two 32 MB simms, and two 8 MB simms, giving me a total of 80 MB of ram. My bios recognizes all ram at bootup, but Linux only recognizes 64 MB. I think this has been addressed here before, something about a line added to lilo.conf. But, I don't remember the solution. Could someone repeat it for me? Thanks very much. To your image line add the string: append=mem=80M --/etc/lilo.conf-- [..] image=/boot/vmlinuz append=mem=80M [..] -- Don't forget to run lilo when you're done. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Lee: Re: smail Solution for Dynamic IP's
On Thu, 26 Feb 1998 22:19:37 EST, Lee Bradshaw wrote: [..] local username had to match the remote username. Check out my web page if you want to see all the effort I've put into HTML so far :^) http://www.mindspring.com/~lee.bradshaw/ And what nice dogs you have. G. If I didn't mention this before, I'm not cc'ing you because ipa.net rejects my from lines as spam. I don't understand why a correct header would be rejected. I'd like to see some details for the basis to this claim, because I use the same address style as you and Daniel. Please tell. I've read the relevant RFC (and more), but some of the terminology is prohibitive to my understanding: - RFC822(STD11) 3.4.6: o Parentheses (( and )) are used to indicate com- ments. o Angle brackets ( and ) are generally used to indicate the presence of a one machine-usable refer- ence (e.g., delimiting mailboxes), possibly including source-routing to the machine. -- What does that bit about one machine-usable reference (e.g., delimiting mailboxes) .. source-routing .. mean? Most importantly are both Joe User [EMAIL PROTECTED] and [EMAIL PROTECTED] (Joe User) in keeping with standards? Is there a more relevant RFC I should be reading? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Lee: Re: smail Solution for Dynamic IP's
On 01 Mar 1998 16:23:02 +1300, Carey Evans wrote: [EMAIL PROTECTED] (David Stern) writes: On Thu, 26 Feb 1998 22:19:37 EST, Lee Bradshaw wrote: [snip] If I didn't mention this before, I'm not cc'ing you because ipa.net rejects my from lines as spam. I don't understand why a correct header would be rejected. I'd like to see some details for the basis to this claim, because I use the same address style as you and Daniel. Please tell. What about the envelope sender? sendmail writes this as: From [EMAIL PROTECTED] Sun Mar 1 16:12:31 There's some strong negative remarks about the use of the From (without the colon) field in the IETF Mailing Headers Draft: --- 3.4 Sender and recipient indication (1) This header field should From (not not standardized never appear in e-mail being followed by a for use in e-mail sent, and should thus not appear colon) in this memo. It is however included, since people often ask about it. This header field is used in the so-called Unix mailbox format, also known as Berkely mailbox format or the MBOX format. This is a format for storing a set of messages in a file. A line beginning with From is used to separate successive messages in such files. This header field will thus appear when you use a text editor to look at a file in the Unix mailbox format. Some mailers also use this format when printing messages on paper. The information in this header field should NOT be used to find an address to which replies to a message are to be sent. (2) Used in Usenet News mail FromRFC 976: 2.4 for transport, to indicate the path or use in Usenet News through which an article has goneFrom when transferred to a new host. (not followed by a colon) Sometimes called From_ header field. By contrast, the Sender: line is standard (though somewhat vague): The person or agent submitting Sender: RFC 822: 4.4.2, the message to the network, if RFC 1123: 5.2.15- other than shown by the From:16, 5.3.7. header field. Should be authenticated, according to RFC 822, but what kind of authentication is not clear. Some implementations expect that the e-mail address used in this field can be used to reach the sender, others do not. See also X-Sender. or something similar at the top of the message. qmail puts: Return-Path: [EMAIL PROTECTED] as RFC821 and RFC822 (4.3.1, 4.4.3) suggest. It appears that the Debian list server renames this to X-Envelope-Sender: before passing the message on, which shows that Lee Bradshaw's envelope sender is (or has been) [EMAIL PROTECTED], and yours (David Stern's) is [EMAIL PROTECTED]. At least localhost will succeed in some DNS lookups. (I had my envelope sender set wrong until recently too. It's difficult to notice.) And it *is* forged by spammers. Some DNS's confirming localhost (127.0.0.1) is probably why I'm able to post here. Seems wrong, but at the moment I'm glad. The Return-Path is where bounced email should go, BTW, which is partly why it gets forged. Have either of you received any bounces lately? I got bounced mailing today: - '[EMAIL PROTECTED]SIZE=2824' sender address target 'localhost' is not a valid e-mail domain. - In case anyone is picking up and missed my original post, I'll reask my questions. I've read the relevant RFC (and more), but some of the terminology is prohibitive to my understanding: - RFC822(STD11) 3.4.6: o Parentheses (( and )) are used to indicate com- ments. o Angle brackets ( and ) are generally used to indicate the presence of a one machine-usable refer- ence (e.g., delimiting mailboxes), possibly including source-routing to the machine. -- What does that bit about one machine-usable reference (e.g., delimiting mailboxes) .. source-routing .. mean? Most importantly are both Joe User [EMAIL PROTECTED] and [EMAIL PROTECTED] (Joe User) in keeping with standards? -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING
Re: Lee: Re: smail Solution for Dynamic IP's
On Sun, 01 Mar 1998 14:34:05 EST, wrote: David Stern wrote: I don't understand why a correct header would be rejected. I'd like to see some details for the basis to this claim, because I use the same address style as you and Daniel. Please tell. This is from /var/spool/smail/msglog/... I broke it into multiple lines Xdefer: [EMAIL PROTECTED] reason: (ERR151) transport smtp: 451 [EMAIL PROTECTED]... Domain must resolve. It is a criminal offense to send unsolicited e-mail to,from,or through this server. The mail seems to be rejected based on the first line: from [EMAIL PROTECTED] not: From: [EMAIL PROTECTED] (Lee Bradshaw) ---IETF Mailing Draft--- - (1) This header field should From (not not standardized never appear in e-mail being followed by a for use in e-mail sent, and should thus not appear colon) in this memo. It is however included, since people often ask about it. not standardized Used to mark header fields defined only in RFC for use in e-mail 1036 for use in Usenet News. These header fields have no standard meaning when appearing in e-mail, some of them may even be used in different ways by different software. When appearing in e-mail, they should be handled with caution. Note that RFC 1036, although generally used as a de-facto standard for Usenet News, is not an official IETF standard or even on the IETF standards track. -- This may or may not be justification for refusal, I can't say for sure. My recommendation would be to avoid using potentially problematic fields, of which From (without the colon) seems to definately qualify. If I set the visible name to mindspring.com, then the from looks like: from [EMAIL PROTECTED] Mindspring inserts a return path: Return-Path: [EMAIL PROTECTED] so my returned mail will go to another user. Regarding lee.bradswhaw going to bradshaw, I thought I read that a dot in this field was not standard and that it may be rewritten if necessary, however I cannot find that now. Again, my impression is that this qualifies as a potentially problematic field entry, and it would be best to avoid the dot. I'm not an authority in this matter, I'm just calling it like I see it. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: diagnosing smail
On Fri, 27 Feb 1998 11:08:26 EST, Daniel Martin wrote: My interpretation is that smail only makes use of the from_field variable if incoming mail has no From: or Sender: fields already; further, if there is a From: field already, then smail inserts a Sender: field using the from_field information only if the pre-existing From: field doesn't match what smail would have written with from_field. IOW: If I'm replying to a message, the header writing will be different than if I send a new message? I'd not fully considered this previously. This would explain why on at least one occasion I got good headers on a new mail, but a reply mail showed faulty headers. I wish I'd thought of this before. Perhaps I'm close to a fix. I'll try getting back to that point and work from there. What is the order of operations between /etc/smail/transports and /etc/smail/config, when modifying, adding, removing header lines, in terms of which goes first and last ? Ok - the idea is that the From: line should be the email address of the person the mail is from, whereas the Sender: is the email address of the account the mail is from, if that's different. (Sender: headers are not required if the From: address also represents the account the mail comes from). [..] IOW: I need the Sender: line because my local hostname is not my isp's? The sender line is good, because it tells where I'm connected when the message was sent, but does it make sense for my mail to be accepted/rejected based on a temporary hostname? This is what my isp told me had occurred. My isp may have misled me about the problem when they said my Sender: line was incorrect. Someone recently mentioned that their From: line was rejected by a specific isp due to the form: [EMAIL PROTECTED] (joe user), which is the same I use. I've read that this form is valid, and that the other usage joe user [EMAIL PROTECTED] is ok to use, but not as well in keeping with the RFC's. (I think I read this in some mh/nmh/exmh docs.) So maybe my mail wasn't being rejected solely on the account of my Sender: line, maybe it was being rejected on the account of the combination of Sender: line *and* the form of my From: line. IOW: my From: line was never looked at previously, because my Sender: line was good, but now that my Sender: line is bad, my From: line is being rejected. This might be a good time to ask: Which form is best? I think there are historic reasons (before from_field was used for Sender: fields) for having it begin with From:. IOW: the from_field code always modifies the From: line? So why did the man page indicate the from_field wrote the From: or the Sender: line? There is one consequence of this which ties in with an earlier point: if the from_field code I inserted is not behaving properly, it could result in my Sender: line being written in a way other than I desired and this could be why my Sender: line is inncorrect. Thanks for the test code; by removing the conditions, I got it to work. The results show that all my variables appear as I've assigned them, so now I just have to figure out where the logic is going awry in my configuration. I wonder what chance there is that MH is causing this turmoil (I switched back to MH from nmh, but at the higher hamm level.) -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
diagnosing smail
Hi,
I'm trying to beat smail down with a stick, am having a little trouble
(actually more than a little). Would somebody *PLease* translate into
English this brief /etc/smail/config snippet and answer three questions?
from_field=From: \
${if def:ident_sender\
[EMAIL PROTECTED]
{$sender${if def:sender_name: ($sender_name)}}}
(My interpretation: look at the From: line (or Sender: line, which?),
and if the sender's identification code (whatever that means) is
non-null (exists), then rewrite that line (which one: From: or Sender:
?) to look like .. here it breaks down, and that extra condition in the
last line confuses me.)
1.) What is the difference between the roles of Sender:, Return-path:,
and From: lines and are each required according to RFC's? (It was my
understanding that there was some contention as to whether the Sender:
line was required, but that it was desirable because it acts as an
envelope, which is a good thing for some MTA's and MUA's. I'm not
clear about the roles of Return-path:, From:, and Sender: or which are
required.)
2.) Why does from_field require it begin From: (as shown in the man
page below) and does from_field rewrite the From: line, or the
Sender: line and how does it decide? (It seems ambiguous.)
8--- man smailconf(5) -8
from_field
type: string
This string will be expanded to form From: or Sender:
fields. The expanded string must begin with From:,
which may be replaced by other strings to form an
actual header field. The default value is:
From: $sender${if def:sender_name: ($sender_name)}
8--8
3.) How can I find out what ident_sender, as well as all other
variables, are assigneded as? (I need to know so I can test what is
happening; e.g.: when I set visible_name to u.washington.edu, and
restart smail and inetd or reboot, there is no change, but I don't know
why. I've tried echoing and mailing the variables, but I must be doing
it wrong.)
I know smail is reading my configuration because smtprewrite is being
implemented, and modifying transports to remove and insert headers
works, yet my attempts to modify /etc/smail/config have mostly been
futile.
--
David Stern
--
http://weber.u.washington.edu/~kotsya
[EMAIL PROTECTED]
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
diagnosing smail (2)
Hi,
This is a slightly shorter version of some questions I asked earlier
regarding smail configuration. I've read all the smail docs available
and there is no definitive smail resource, so I'd really appreciate it
if someone would *PLease* take a few moments to reply.
1.) How can I find out what ident_sender, as well as all other
variables, are assigneded as? (I need to know so I can test what is
happening; e.g.: when I set visible_name to u.washington.edu, and
restart smail and inetd or reboot, there is no change, but I don't know
why. I've tried echoing and mailing the variables, but I must be doing
it wrong.)
2.) Why does from_field require it begin From: (as shown in the man
page below) and does from_field rewrite the From: line, or the
Sender: line and how does it decide? (It seems ambiguous.)
8--- man smailconf(5) -8
from_field
type: string
This string will be expanded to form From: or Sender:
fields. The expanded string must begin with From:,
which may be replaced by other strings to form an
actual header field. The default value is:
From: $sender${if def:sender_name: ($sender_name)}
8--8
3.) What is the difference between the roles of Sender:, Return-path:,
and From: lines and are each required according to RFC's? (It was my
understanding that there was some contention as to whether the Sender:
line was required, but that it was desirable because it acts as an
envelope, which is a good thing for some MTA's and MUA's. I'm not
clear about the roles of Return-path:, From:, and Sender: or which are
required.)
Any help would be greatly appreciated.
--
David Stern
--
http://weber.u.washington.edu/~kotsya
[EMAIL PROTECTED]
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: X install/config query
On Fri, 27 Feb 1998 09:12:35 MST, wrote: Hello: Hello I recently installed Debian 1.3 on a '386, but apparently did not complete the process correctly. X was not configured (although everything else that I have tried seems to be). If dselect or dpkg says your xserver isn't configured, you might try running through [C]onfigure once or twice in dselect. I cannot find 'xf86config' or 'XF86Setup' on the system. Because configuration is incomplete, this could mean either it's not there, or you're not seeing it. You can always run the xserver post install script in /var/lib/dpkg/info/ and confirm that the files in the xserver list file, also in /var/run/dpkg/info/ exist. (There's no sense running configure if the files aren't there, and I don't know what kind of problem you encountered.) Should I remove X and re-install, or is there another way? Sometimes dselect is quirky and it takes a couple tries, I've found. If that doesn't work, I'd remove my X-server, then confirm that no outstanding conflicts or dependencies exist in dselect by running though installation in dselect, and then run through installation again in dselect, choosing your xserver. I think that XF86Setup (nicer than xf86config), is only in the VGA xserver, reason being that this is supposed to be a good default. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: root access and dselect | ftp
On 25 Feb 1998 22:47:23 +0100, Martin Bialasinski wrote: [EMAIL PROTECTED] (David Stern) writes: [..] This is why I was asking about dselect | ftp, because if I'm root, and I'm running ftp via dselect, then isn't this exactly what you're telling me not to do? Yes, but there is no known way to force the ftpclient to do such things. The client doesn't accept any commands and any data it gets is, well, data, so it is not executed, just written to disk. Initially I thought that ftpd accepted commands, but now that I think about it a little more, I suppose it accepts data. (at least for ftp clients, running ftpd in server mode on the net might be a different animal) Running an ftp client as root seems to be an exception to the rule about not running as root. Thanks for the clarification. :-) p.s.: I won't make personal replies until my headers are up to snuff, which I'm working on. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Question
On Thu, 26 Feb 1998 09:45:09 PST, Erik Rodr guez wrote: I have Windows NT Workstation and Debian Linux installed on my computer, usually i download files from Internet in Windows 95 (at the office), today i downloaded the java-linux package but i have it in Win 95, how can i put it in my home's linux machine, the problem is that my linux isn't connected to a network, but it's installed in the same PC as Windows NT. Your question is not very specific. I can think of a few solutions: 1.) install dialup networking on your home computer 2.) use a portable drive (zip, jaz, syquest, LS-120, tape, ..) 3.) sneakernet (split the file up onto multiple floppies) Since your next mail has the same subject as this one, I'll add that dselect can be a little bit finicky at times, so you may want to try to run through the installation a few times to see if some of the errors go away. Installation is thoroughly covered in several online documents, so I'd recommend you get plugged into the resources which are available to you. Installing Debian ftp://ftp.us.debian.org/debian/stable/disks-i386/current/install.html Debian Linux User's Guide http://www.linuxpress.com/001001.htm Debian Installation and Getting Started http://www.ssc.com/lg/issue15/debian.html p.s.: I will not make personal replies until my headers are fixed. -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
lean indexed navigation tool for system help
Hi, :-) Someone using a lesser linux distro is asking me if there's any kind of a lean navigation tool which indexes the typical Linux help facilities built into his box such as man and info pages, HOWTO's, /usr/doc/*.gz, and so on. My only idea was that something like that might be possible using a search engine (ht-dig), a web-server (apache), and a browser (Lynx or Netscape) setup with mime-types for the various doc formats and locations. However, he would like a leaner, ready-made utility. Do any of you less mortals know of anything that would fit this description? p.s.: don't look at my headers yet. :-) -- David Stern -- http://weber.u.washington.edu/~kotsya [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new hamm nmh breaks header rewriting, isp becomes irate
On Sun, 22 Feb 1998 17:43:39 EST, wrote:
[EMAIL PROTECTED] (David Stern) writes:
The most likely explanation is that nmh has started adding Sender:
lines; this is in general a good thing - we just need to be careful to
take them out or rewrite them nicely on the way out. It may be that
some machines are rejecting your mail for having sender lines with two
@ signs in them. In that case, a kludge to fix it is to change the
line that adds the sender stuff to:
insert_header=Sender: \
${if def:ident_sender \
[EMAIL PROTECTED]@$visible_name}}
This, combined with the remove_header line above, should get things
back to the way they were.
This idea sound effective, but I had problems getting smail to take
this, probably because I'm not putting it in the correct place -- where
exactly does this go?
Hmm... Sender headers really shouldn't be rewritten like this if they
already exist... Perhaps something like:
from_field=From: \
${if def:ident_sender \
[EMAIL PROTECTED] \
{$sender${if def:sender_name: ($sender_name)}}}
in /etc/smail/config and then nothing dealing with Sender: headers in
the transports file (neither adding or removing) would be better.
This sounds more artful, and I think I sort of got this to work (it's
as good as it was before, at least -- more on this in a second).
There's a double quote missing from the end. Thanks, Daniel.
You may also want to ensure that the visible_name used is something
other than localhost, which is what it appears to be set to.
Unfortunately, the only way I've found to do that (without having a
name registered with .dyn.ml.org) is to rewrite /etc/smail/config each
time ip-up is called. (There's one way of doing that on my webpage
http://www.math.jhu.edu/~martind/mybox.html - after I wrote that page
I figured out a cleaner method using m4)
I thought I only needed that set visible_hostname to my dynamically
assigned IPA if I needed to be able to be contacted directly there. I
see one potential conflict with setting my visible_hostname to my
dynamically assigned IPA, and that would be depending on how smart my
smarthost is, bounced mails, as sometimes occur for reasons other than
my Sender: line, may be bounced to my isp after I disconnect, and right
now, that would be a bad thing. I know that the visible_hostname,
mx[1-4].u.washington.edu, set by allowing the default set at runtime
works for mail delivery, because I've tried it, but the dynamically
assigned IPA I'd get would be much different (something like
cs_student_XXX.washington.edu), and I'm afraid to test that, because my
isp is irate with me right now. What exactly do the RFC's say about
this?
(One of these days I'm just going to get fed up and write a mailer
designed for dialup systems which need to rewrite headers on the way
out and may well have no consistent name - the wonders of free
software that I won't have to start from scratch...)
I don't know why there aren't provisions made for this in the
traditional MTA's, because dialup networking is probably one of the
most common types of internet connections now. This should be standard.
I really appreciate your help. It'd have taken me eons to figure this
out alone. Thanks, Daniel.
--
David Stern
--
http://weber.u.washington.edu/~kotsya
[EMAIL PROTECTED]
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
