Re: writing to /dev/stderr
echo test | gawk '{print $0 /dev/fd/2;}'
gawk: cmd. line:1: (FILENAME=- FNR=1) fatal: can't redirect to
`/dev/fd/2' (Permission denied)
There's a kernel bug operating here, I think. If you chmod 666 /dev/fd/2
(from the same console, of course) it shows no change in its mode using
ls, but you will probably be able to write it. I was able to verify
that I could chmod it to 000 and would get the same no-permission message,
but the mode did not appear to be 000 using ls.
Bruce
Re: writing to /dev/stderr
I looked at this a little more. It seems that the object on the other side of /dev/fd/X isn't always one that you have permissions on _after_ it's been opened. ls -l /dev/fd/2 lrwx-- 1 brucebruce 64 Aug 30 16:16 /dev/fd/2 - [0301]:7060 This means device 3, 1, inode 7060. ls -li `tty` 7060 crw-rw-rw- 1 root tty4, 193 Aug 30 16:16 /dev/ttyp1 Note the 7060. ls -li /dev/hda1 6237 brw-rw 1 root disk 3, 1 Dec 31 1969 /dev/hda1 That's the root partition. Note the 3, 1 It's the permissions of /dev/ttyp1 that apply here, not those of /dev/fd/2 . Bruce
Re: writing to /dev/stderr
On Fri, 30 Aug 1996, Alex Romosan wrote: the links were okay, the problem is elsewhere. if i telnet into the machine and log in as alex gawk works fine. what i usually do is to rlogin from an sgi to a different account and then su to alex. under these circumstances gawk fails with permission denied. i have no idea if this is a bug or not. Ok, now it's clear. The difference between those two methods is who owns the controlling terminal. When you su, the old user still owns it, so the new one can't write directly to the tty. And /dev/fd/2 is implemented as a link to the tty: lrwx-- 1 gmaorgmaor 64 Aug 30 19:38 /dev/fd/2 - [0802]:556 That means device 8,2 inode 556. 8,2 is my root partition, and 556 is ttyp4. 556 crw--w 1 gmaortty4, 196 Aug 30 19:39 ttyp4 So if I su to someone else, I still own ttyp4 and the new user can't write to ttyp4. As is apparent, I was wrong about /dev/stderr being implemented internally by gawk, at least under linux. Using strace and from the above, it's pretty obvious that gawk really is opening /dev/stderr, If it were just writing to fd 2, it would work. This is one case where using fd 2 is not the same as writing to /dev/stderr. The solution is to switch user id's by doing 'exec login'. You'll transfer the controlling terminal also then. Guy
Re: writing to /dev/stderr
On Fri, 30 Aug 1996, Bruce Perens wrote: It's the permissions of /dev/ttyp1 that apply here, not those of /dev/fd/2 . Just like in a symbolic link. Guy
Experimental DHCPD package
The only thing you have to do is to customize the
/etc/dhcpd.conf file a bit to get it going on your machine. (That is if it
works for you)
Available at
ftp://ftp.fuller.edu/Linux/debian/dhcpd*
Please report back to me your experiences. If some people run it
successfully then I will submit it to be included in the Debian
distribution.
{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}
{}Snail Mail: FTS Box 466, 135 N.Oakland Ave, Pasadena, CA 91182 {}
{}FISH Internet System Administrator at Fuller Theological Seminary {}
{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}
PGP Public Key = FB 9B 31 21 04 1E 3A 33 C7 62 2F C0 CD 81 CA B5
Re: XFree86 3.1.2F plans?
I agree. I would very much like to be running the F beta in its entirety, and with Debian I do not know how to go about doing that and being positive I am not breaking anything in the process. I realize that the owner of this package does not plan to release any of the beta versions as a Debian package while there is no source code, but I for one would VERY much appreciate some sort of instruction on what all is neccessary in order to take out the stock Debian XFree and replace it with one of the beta versions without breaking anything in the process. I don't care if the process is manual or automatic, I just want to be able to do it. Under Slackware you just had to rip out the old and put in the new, and that was basically it, but under Debian, with the Debianized XFree86, not even knowing the extent really of how it differs from the standard XFree86, I don't want to experiment with it. Perhaps just using dpkg to remove XFree86 without removing any of the packages depend on it, and then installing the beta version the normal way by just untarring it? That will mean that the conf files and stuff will be in the non-debianized locations and all, but what I don't know is whether or not this would break anything else. Insight on this matter would be very much appreciated. Every time I've brought this up in the past everybody just says to use the server from the beta package, which I already am, but I would like to go beyond that, and run the whole beta package. It seems like in this respect, Debian is limiting my possibilities instead of making it easier to do what I want with my system. Shawn Asmussen On Fri, 30 Aug 1996, Randall Shutt wrote: On Aug 29, 8:15pm, Michael Alan Dorman wrote: Subject: Re: XFree86 3.1.2F plans? If one needs a server that is not included (or the version included is inadequate) in the current non-Beta XFree86, you need only get the _server_ (not the support files), unpack it into /usr/local, and modify the first line in /etc/X11/Xserver to point to it. Well, what of those that want to try the whole beta? Its so radically different, everything that I had in /etc/X11 got stomped on my /etc/X11 is no more. Alot of config stuff is different. So what Im asking is, how best could I have done this as to not completely fuck up the debian package database (I was unable to dpkg -r alot of stuff due to fear of also removing subsidiary packages that depended on X11). So, now my main question is, when they DO come out with a .deb or a .tar.gz with debian rules, how will I best go about not totally stomping all over everything again, and also not having unaccounted for files? Thanks -- === Randall Shutt RaveNet Systems, Inc. ===
Using dpkg to install Debian?
Hi there. I have a couple of questions, the first of which is: is it possible to install *components* of Debian without having to totally blast away all of the Slackware garbage I have now? The version I have is old enough that it's due upgrading anyway, and I'd like to switch to Debian piecemeal 'cause of the package stuff being so much easier. But I don't want to overwrite everything, because I don't want to break the stuff I have right now and because I don't want to spend thirty hours on the net downloading everything. :) I assumed that the logical approach would be to download the source to dpkg and dpkg-ftp, install them, and then install some of the more critical pacakges via ftp. After *finally* locating dpkg-ftp, I recompiled everything on my Slackware machine. (This might be stupid, I don't know. Does the Debian elf format differ from the Slackware elf format? i.e., did I *have* to recompile, or can I just grab the binary-i386 stuff?) configure ran fine except for one error: it was unable to determine my architecture. Now, this is kind of strange. When I accidentally tried to run it on a sun4c, it claimed an error with dpkg on line one of configure (?), looked like a shell error, but on my Linux machine it just gave me something about CC not being found. Anyway, I tried make anyway, which of course didn't work. But I edited config.h and manually set ARCHITECTURE to i586. (Should this be i386?) After that, compile worked fine for everything except dselect, which died on a bunch of ncurses include files. After doing some symlinking and changing some strange functions to ncurses ones, it compiled but wouldn't run. (It'd run, but nothing happened. Looks like I broke ncurses in the process of getting dselect working.) Anyway, dpkg seems to work mostly fine once I made a /usr/doc/copyrights directory for the FSF GPL. Soo... I tried running dpkg -i on dpkg-ftp_whatever-version-it-is.deb, and it gave me the following error: dpkg: failed to open package into file `/var/lib/dpkg/status' for reading: No such file or directory I'm pretty much a Debian newbie, so I really don't know what I should do. Hell, I don't even know if dpkg -i installs stuff (although the fact that it's the same as dpkg --install would seem to imply that), not for sure. I'm running it as root. Doing a dpkg --update-avail Packages works fine (once I figured out that you're supposed to gunzip it... drr) and gives me a dselect-like output. I've been over the FAQ, both the old and the new, at least thirteen times and saw nothing of relevance. Do I need to do a 'touch /var/lib/dpkg/status' or something? What's the file *for*? Thanks, and I'm sorry this got so long. I know I'm asking for it by not just backing up all the stuff on my hard drive, but downloading 60 megabytes of software really just isn't an option, and neither is spending the money to get Debian on CD-ROM. (I'm broke.) -- Barid Bel Medar [EMAIL PROTECTED] Knights of the CosmosShayol Ghul Resort and Health Spa -- I am returning this otherwise good typing paper to you because someone has printed gibberish all over it and put your name at the top. - English Professor, Ohio University --
Re: UPDATE: deb-view.el 1.2: emacs tool for browsing deb files!
Mark Eichin [EMAIL PROTECTED] writes: w3 can be found via anonymous ftp to ftp.cs.indiana.edu:/pub/elisp/w3. More usefully, w3 can be found in the binary-i386/net/w3-el_2.2.25-4.deb on any debian mirror... Yeah, but those 2.2.x w3's are way slower than the 2.3 now 3.0 betas, which are very usable. There's a new html parser. It's a dramatic difference, I really recommend it, well worth what betaness there is. In the 'betas' subdirectory of the above path; there may be a symlink to newest version in the above directory also, don't recall. -- Ed Donovan [EMAIL PROTECTED]
Re: UPDATE: deb-view.el 1.2: emacs tool for browsing deb files!
Ed Donovan [EMAIL PROTECTED] (me) writes: Mark Eichin [EMAIL PROTECTED] writes: More usefully, w3 can be found in the binary-i386/net/w3-el_2.2.25-4.deb on any debian mirror... Yeah, but those 2.2.x w3's are way slower than the 2.3 now 3.0 betas, which are very usable. There's a new html parser. It's a dramatic difference, I really recommend it, well worth what betaness there is. In the 'betas' subdirectory of the above path; there may be a symlink to newest version in the above directory also, don't recall. Well, wait, I didn't add: there is also a 2.2.26 release version, which I'm pretty_ includes the fast new parser, though not this and that spiffiness, blah, blah... Raul, if you want to, great...if you don't have the time, maybe I'll give it a shot. Thanks, -- Ed Donovan [EMAIL PROTECTED]
Re: Using dpkg to install Debian?
From: Barid Bel Medar [EMAIL PROTECTED] Is it possible to install *components* of Debian without having to totally blast away all of the Slackware garbage I have now? The version I have is old enough that it's due upgrading anyway, and I'd like to switch to Debian piecemeal 'cause of the package stuff being so much easier. But I don't want to overwrite everything, because I don't want to break the stuff I have right now and because I don't want to spend thirty hours on the net downloading everything. :) It is much more trouble than it is worth to install the system piecemal over Slackware. We _can't_ ever support that, even if we wanted to. Sorry. More stuff will go wrong than would if you did a cold install, and you will end up spending much more time on the project. If you can add disk or have an extra partition available, we suggest you install Debian on one partition and have Slackware on another so that you can copy things from your old partition. If 30 hours on the net isn't do-able, you might patronize one of the CD vendors and spend $30 instead. See our web page www.debian.org for a few vendors. Thanks Bruce
Re: Using dpkg to install Debian?
Hi Barid -- I assumed that the logical approach would be to download the source to dpkg and dpkg-ftp, install them, and then install some of the more critical pacakges via ftp. Well, the better thing to do is to -- download boot1440.bin, root.bin, base14-1.bin, base14-2.bin, and base14-3.bin (or the 1220 equivalents if you use a 5.25 floppy) from the buzz-fixed/disks-i386/current/ directory, then -- copy those files to floppies using rawrite (or dd since you already have Slackware going), then -- boot with the boot floppy, follow directions in the menu interface to set up a new partition, and it will install a Debian system (including dpkg and dpkg-ftp), etc. for you. I've been over the FAQ, both the old and the new, at least thirteen times and saw nothing of relevance. Do I need to do a 'touch /var/lib/dpkg/status' or something? What's the file *for*? (That means we need to fix the FAQ somehow.) The file /var/lib/dpkg/status keeps track of what you've got on your system. It has entries like this: Package: vim Status: install ok installed Priority: optional Section: editors Maintainer: D.J. Gregor [EMAIL PROTECTED] Version: 3.0-5 Description: VI iMproved - enhanced vi editor Vim is an almost compatible version of the UNIX editor vi. Only the 'Q' command is missing (you don't need it). Many new features have been added: multi level undo, command line history, filename completion, block operations, etc. See /usr/doc/vim/difference.doc for more information. The critical line is the Status line. If you were to execute the command dpkg --status vim you'd see that stanza printed out on your screen. Hope that's some help. Susan Kleinmann
Re: Making filesystems at installation time
Bill Wohler writes (Making filesystems at installation time): ... The idea with all this installation stuff is to avoid having user actions interspersed with automatic actions that take a while. This also goes for dpkg -GROEB --configure--all the packages that require user intervention for setup should be postponed until last. That can't be done, because (a) dpkg doesn't know and (b) the configure order is largely determined by the dependencies. dpkg already unpacks packages first and then configures them all later, and the unpack is (almost) entirely noninteractive. Ian.
Re: Debian Color-ls Command?
With the integration of color-ls directly into the fileutils package, a few things have changed. dircolors no longer sets up aliases or shell scripts to colorize ls, dir, and vdir. Here is an excerpt from a .bashrc which sets up aliases after running dircolors: # set up color-ls eval `dircolors /home/syrus/.dir_colors` alias d='ls -F --color=auto' alias v='ls -l --color=auto' alias vdir='ls -l --color=auto' alias dir='ls -F --color=auto' Note that color=tty has been changed to color=auto. See the documentation for other change information. Cheers. Syrus. Hi Syrus, In fact the above didn't work for me either--dumped core the first time I used 'ls' after that! Had a hell of a time trying to figure it out.In fact the new way of doing things (at least in my case, and by new I mean Debian 1.1) isn't to rely on /etc/DIR_COLORS, as color-ls and dircolors don't seem to read it! What I had to do was run dircolors, monitor its output, then copy this into my profile (/home/cw/.bash_profile). This is what finally did the trick: *** Excerpt from .bash_profile *** # set up color-ls LS_COLORS='di=1;34:bd=40;33;1:ln=1;36:cd=40;33;1:ex=1;32:*.tar=1;31:*.tgz=1;31:*.gz=1;31:' export LS_COLORS; LS_OPTIONS='--8bit --color=tty -F'; export LS_OPTIONS; alias ls='/usr/bin/color-ls $LS_OPTIONS '; alias dir='/usr/bin/color-ls $LS_OPTIONS --format=vertical'; alias vdir='/usr/bin/color-ls $LS_OPTIONS --format=long'; alias d=dir; alias v=vdir; Of course, for all I know this could the *old* way of doing things, but it works! Now I'm happy. Many thanks to all who responded. Cheers, Chris
Re: XFree86 3.1.2F plans?
On Fri, 30 Aug 1996, Shawn Asmussen wrote: I agree. I would very much like to be running the F beta in its entirety, and with Debian I do not know how to go about doing that and being positive I am not breaking anything in the process. I realize that the owner of this package does not plan to release any of the beta versions as a Debian package while there is no source code, but I for one would VERY much appreciate some sort of instruction on what all is neccessary in order to take out the stock Debian XFree and replace it with one of the beta versions without breaking anything in the process. Okay. Here's how I got X 3.1.2[D|E|F] running. The background: The machines I was given to work with have Mach64 chipsets and the hardware appears to be so blasted _new_ I couldn't get the .deb X to work. Ditto the X11R6 implementation of XFree86. I _had_ to go with the betas, Debian or not. This might make some people cringe, but I didn't take out the Debian XFree--I didn't want to be forcing overrides of (what the machines thinks would be) dependency problems for the next n+1 months--with it installed but more than a little broken, I pulled the D beta stuff into /usr/X11R6 and installed it right on top of the Debian X. Because it was after D's expiration date when I did this, I then I pulled down the E stuff and upgraded. Yesterday and Thursday, I pulled the F beta stuff down and upgraded again. I'll continue to do this with each successive beta, until the gamma is released. This worked, first time. To date, every time. Knock on monitor. It's wildy wrong from the Debian package database's point of view, but that's life. It's as messy as hell, but dpkg's ability notwithstanding, sooner or later, hacks must be perpetuated in order to get things to work and messy things will be effected. Things are going to become INTERESTING when the .deb version of X11R6.1 has been released and its time to install it. Will I find files that aren't accounted for by the .deb version? Will I find other random weirdness? I'll deal with that when I run across it. I realize that the question of how do I install the betas without making a mess of the package database is as important as how do I get the betas running to the last two or three people who've asked. I also realize that this message doesn't even attempt to answer it, because I don't have experience with the next phase of it (installing the .deb over the betas), and I apologize for this. David [EMAIL PROTECTED] Office: 3503 WeH, x86720 MTFBWY
Do you use SLIP or a variant with Debian?
If so, I'd appreciate a short note from you. I'd like to know if you use SLIP because PPP is unavailable, more expensive, or otherwise inconvenient. Thanks Mike
hardware needed for real terminals..or consoles
one thing I hope to do when I get acquainted(future,hopefully near )is to open consoles so mt family has, in effect simultaneous access to the computer so I can make upgrades nto one computer rather than buy and maintain several since hardware tends not to be cheap, I'm trying to think(and save up) ahead of time when that day comes what is needed for each terminal?
Re: Need Help
Hi! Have you succeeded in installing Debian? Yes and no. Following advice from Syrus, I tried different sets of disks. The fixed set of disks seem to have worked, but I think I may need to do FIPS again since I did not get/make a swap partition. I need to get some apps running and see what happens. Maybe the next time I send something to the group it will be from Debian. :) Matt
caffiene-free netscape
I have installed Netscape 3.0 (had to pull the deb package from the incoming directory since for some reason the packages file still has the last beta as the current version) on my debian system. The browser works fine, but I can't get any java applets to work. I have the setting set right and used the package setup that Brian put together, but something still isn't working. I don't receive any errors when I aim at a page that I know has an applet on it, it just doesn't work. The applet space is blank. I am wondering if there is anything else I would need to do to get java to kick in. I am not ready to start developing applets yet, I just want to have them work. I have the 2.0.6 kernel installed (compiled myself). Any ideas on where to look for settings that aren't right? -- David Morris [EMAIL PROTECTED] --- The fundamental question in life is whether we live it from the outside in or from the inside out. - Fritz Perls (founder, Gestalt therapy) --- Home Page: http://www.netins.net/showcase/dmorris/ PGP Public Key available there or via Finger. ---
Re: hardware needed for real terminals..or consoles
one thing I hope to do when I get acquainted(future,hopefully near )is to open consoles so mt family has, in effect simultaneous access to the computer so I can make upgrades nto one computer rather than buy and maintain several since hardware tends not to be cheap, I'm trying to think(and save up) ahead of time when that day comes what is needed for each terminal? A serial port. A serial null-modem cable. A terminal. A free plug ;-). Seriously, that is about all you need, hardware wise. Out of them, the hardest to come by would be the serial ports and the terminals. I don't know where to go to get dedicated terminals, but usually some are available in the used market. The serial ports are a potential problem. Standard PC hardware limits you to two or four serial ports, unless something special is done. There are plenty of solutions, most providing 4 or 8 extra serial ports. If this is all you need, then that would be fine. The cabling is simple, and you might even be able to buy pre-made null modem cables. However, this solution, while fairly traditional, does have certain limitations. The terminals will be limited to text. X will be impossible. If this is acceptable to you, then this is probably the least expensive option. If you want X-like capabilities, then you need to go a more complicated route, a route that most likely involves ethernet. I'll let others deal with that problem. -- Buddha Buck [EMAIL PROTECTED] Just as the strength of the Internet is chaos, so the strength of our liberty depends upon the chaos and cacaphony of the unfettered speech the First Amendment protects. -- A.L.A. v. U.S. Dept. of Justice
debian user in SW Michigan(Kalamazoo-St.Joe counties)
I get the feeling that I am obtusely missing some important underlying point here that only tete a tete communication will solve. So if anyone in Kalamazoo-Three Rivers vector is willing to help me I'd be ever so grateful ..Mike List
/dev/audio /dev/dsp Device or resource busy ???
Hello DEBIAN user/programmers, After kernel v2.0.0 I have been unable to produce any sound on my SB16 sound card. Is there a known bug (up until kernel 2.0.15) in the SB16 code of USS that I missed during my vacations? The sparing comments in the source point to an IRQ or DMA conflict when one gets constant Device or Resource busy mesages on each: cat blabla.au /dev/audioor cat uuhuu.wav /dev/dsp But I haven't changed anything in the configuration of my hardware, nor my kernel parameters... Sorry but I didn't encounter any hint's in my /var/adm/messages eighter. The USS does not permit compilation with DEBUGING on, does it? Thankfull for any hints or suggestions, regards, Stoyan -- Stoyan Kenderov/ phone: +49 721 9652 220 NTG Netzwerk und Telematic GmbH \/ fax: +49 721 9652 210 Vincenz-Priessnitz-Str. 3/\ LINK email: [EMAIL PROTECTED] D-76131 Karlsruhe, Germany /___ http://www.xlink.net/~kenderov Opinions stated herein are my own, not necessarily my employer's [finger me for my PGP public key]
inconsistency/confusion in aout-svgalib of Debian 1.1
I'm not sure this is the right place to send this bug report... There's an inconsistency in Debian 1.1. When the aout-svgalib package is installed, it puts the libary files in /usr/i486-linuxaout/lib, but ld.so is not configured to look there for libraries, so the aout svgalibs are unusable. There seem to be two simple ways to fix the problem: * a one-line addition to /etc/ld.so.conf * installing the libraries etc in /usr/lib/i486-linuxaout instead I'm not sure which way is The One True Way. Sebastian
Re: caffiene-free netscape
On Sat, 31 Aug 1996, David Morris wrote: The browser works fine, but I can't get any java applets to work. I have the setting set right and used the package setup that Brian put together, but something still isn't working. I don't receive any errors when I aim at a page that I know has an applet on it, it just doesn't work. The applet space is blank. A possibility is that you're not waiting long enough. In my experience, it looks like the page is loaded (Netscape will indicate that the document is loaded) and there is just a blank where the applet is supposed to run, but in reality, either the applet is still loading or it is just taking a long time to initialize or whatever. The delay will probably be longer if you do not have a lot of RAM as the applets seem to use lots of memory. Gerry
Debian Installation Boot Program missing something
My disk controller seems to be wanting to be reset, but the
Debian install boot doesn't do this. As a consequence, it just
keeps repeating
hda: hda: Status error: status=0x01 { Error }
hda: Status error: error=0x04 { DriveStatusError }
hda: Drive not ready for command.
The Slackware install boot on the other hand, does the following.
hda: hda: Status error: status=0x01 { Error }
hda: Status error: error=0x04 { DriveStatusError }
hda: Drive not ready for command.
[.. this error msg is repeated thrice.. ]
ide0: reset success
hda1 hda2 hda3
VFS: Insert root floppy disk to be loaded into ramdisk and press ENTER.
I guess no Debian for me for now.
-bhaskar
Re: UPDATE: deb-view.el 1.2: emacs tool for browsing deb files!
Well, how many times can I follow myself up? Let me know if I get near a record. Ed Donovan [EMAIL PROTECTED] writes: Well, wait, I didn't add: there is also a 2.2.26 release version, which I'm pretty_ includes the fast new parser, though not this and that spiffiness, blah, blah... Uh, I may or may not be pretty, but that should have said ...I'm pretty sure... with little underlines which I won't try again. Thank you for your support. -- Ed Donovan [EMAIL PROTECTED]
trouble patching kernel
I've been running the original 1.1 distribution for a while, and I'm finally getting around to patching the kernel up to the current level. This isn't going quite as smoothly as I expected. I fetched all the patches and tried to apply them and experienced: /usr/src # linux/scripts/patch-kernel linux linux/Patches Current kernel version is 2.0.0 Applying patch-2.0.1.gz... 1 out of 1 hunks failed--saving rejects to drivers/sound/Config.in.rej failed. Clean up yourself. What did I do wrong? -- Bill Roman ([EMAIL PROTECTED] / [EMAIL PROTECTED]) running linux
