Re: Iostream.h
Kevin, > #include and the compile fails, does anyone know where I can > get this H file? what packages is it in ... is this a C++ Header file of > a C Header File? Try /usr/include/g++, /usr/include, /usr/lib/g++-include. You will need a copy of g++. Mail me direct, if you need a URL for gcc-2.7.2 (which includes g++), or glib++. 73, -- AA9AU Morse code HaiKu dadidadit dadadah dadidit dit dahdididit dididadit didit dadit dit (rot13) pbqr o svar FANUC Robotics North America, Inc. [EMAIL PROTECTED] Voice: 1-248-377-7469 Fax: 1-248-377-7363 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Recording sound
Peter, > What package do people use to record .au audio files? I've had good service from xwave-0.6 and xmix available from most of the sunsite and other archives. (Please mail me direct for a URL if necessary.) xmix controls the SoundBlaster card, and xwave records, plays and displays the results. I use octave-2.0 for signal processing. Reporting, -- Robert Meier "COBOL is here for the long run... [O]n the Starship Enterprise make no mistake, the program that calculates their pay will be written in COBOL..." - Jerome Jahnke FANUC Robotics North America, Inc. [EMAIL PROTECTED] Voice: 1-248-377-7469 Fax: 1-248-377-7363 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: smail config
Alex, > How do I tell smail, that a single user shouldn't be allowed > to send/receive email? If the user is a fictitious user, you can simply put a .forward file in their home directory (see smail(8), smail(5)). If the user has no home directory, you can place an entry in /etc/aliases (see aliases(4)). Many years ago, after all else failed to curb an abusive user, the rules in /etc/sendmail.cf (see smail(5)) were modified such that all his transmissions were redirected to censors. Shortly after that, his account was revoked for other reasons. Hopefully helpful, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian in a lab
Christoph, > ..., I usually install [packages] ... in a directory > /usr/local/packages/ ... > I use a script to move files into the correct directories. > The appeal of such an approach is [ability to] use any packages ... without > [installing] anything locally except a set of well-defined links[,] > and without using up local storage or worrying about updates. ... > this is a way of providing software to people without having to be root > on their machines and thus without being responsible for their mistakes... > I found that this works extremely well with most packages. Is anyone thinking > of adding a concept like this to standard Debian as an option? Yes! Ray Ingles and myself are currently defining this concept and would appreciate your help. Please see our prior mail (sent 4/15/97) "Re: DEITY TEAM -- REQUEST FOR FUNCTIONALITY and COMMENTS". Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: DEITY TEAM -- REQUEST FOR FUNCTIONALITY and COMMENTS
Peter, Thank you for request for ideas and desires regarding the next improvement to the debian package management system. 1. Scripts provided by the package writer should only have access to files and directories specifically approved by the installer. 2. Most packages do not need to alter existing system directories or files, and should be installed and tested by an unprivelaged user (specified by the installer) in a directory chosen by the installer, and under which package scripts can create and modify files. 3. After testing, the installer should use ln -s, ln, or cp (as chosen by the installer) to integrate the package executables and files into the system. Ray Ingles and I, have spent some time discussing improvements to dpkg/dselect to permit users to take advantage of its dependency tracking without the security vulnerability entailed in always running it as root. The following is a first draft of a processing model (similar to the ISO network model) that hopes to provide the following: 1. Host selectable security - the installer chooses what level of trust (unprivelaged, privelaged, root) to grant to the package scripts. 2. Host testing - before the package is seen by other users, the installer can test the package 3. Portability - package writer can assume a single (or small number) of directories in which to create, modify, compile, configure, files and executables, independent of the platform or host cut here * Project: debian File:RFC: dpkg target model Author: Raymond A. Ingles Dr. Robert J. Meier, Jr. History: 97-04-03 -rjm- file creation * Goals ** ease of use The package provider and the installation process should automate as much of the installation and removal as feasible for ease of use. All operations should have defaults to support ease of use. ** security As far as possible, malicious or buggy package installation should not endanger existing installations. All default operations should be defined by the install procedure so as not to endanger existing installations. All package-suggested operation parameters must be individually approvable by the human installer. Successful or unsuccessful installation is completely reversible. ** flexibility As far as possible, package installation should be configurable by the host to meet individual user needs and concerns. As far as possible, package installation should be configurable by the host to meet individual package needs and concerns. All install operation parameters should be selectable by the installer. All install operation parameters should be suggestible by the package. ** repeatability As far as possible, package installation should produce the same behavior on different hosts (e.g. the package provider and the user). By default, installation will be done under a single host-selected directory with an image equivalent on the user host to that to the package provider host. * For design purposes, installation is divided into the following phases. ** (Template) Each phase needs to answer the provide answers to each of the following questions. The answers must express the minimum/default/maximum supplied by/required from the package/host. *** System privileges *** Host information *** Package information *** Intended results *** Prior assumptions *** Actions *** Validation *** Customization ** Download *** System privileges Minimum supplied by host: write a host-specified file as $DOWNLOADER. Default supplied by host: write a host-specified file as $DOWNLOADER. Maximum supplied by host: write host-specified files as $DOWNLOADER *** Host information Minimum supplied by host: $PACKAGEROOT Default supplied by host: $PACKAGEROOT Maximum supplied by package: filenames *** Package information Minimum supplied by package: number and description of required files and directories. Default supplied by package: number and description of required files (1) and directories (1) *** Intended results Minimum supplied by host: transfer the package to local file system Default supplied by host: transfer the package to local file system Maximum supplied by host: transfer the package to local file system Minimum supplied by package: from package file Default supplied by package: ftp, cd-read, floppy-read Maximum supplied by package: from net, cd, floppies, tape, etc. *** Prior assumptions Minimum supplied by package: the complete package is transferrable as a single file Default supplied by package: the complete package is a compressed tar file *** Actions Minimum supplied by host: Create a specified file in (a directory chosen by host) writable by $DOWNLOADER. Default supplied by host: Cr
Re: RPM
Bruce, > I would like to see a proposal for reduced-privilege installation using > dpkg/dselect. Feel free to write one and make a test implementation. This is a wonderful idea. The convenience and dependency management of dpkg with the reliability, security, tolerance, and flexibility of traditional package management would be the better of both worlds. I added this to my list of projects to do. If someone else doesn't do it before I finish other projects, a friend and I will. In the meantime, I and a friend would be interested in hearing from others interested (please mail direct rather than use the list bandwidth). Can you suggest any design documents for dpkg (in addition to the source and the /usr/doc/* files). Thank you, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: RPM
Bruce, > Note that Red Hat, Caldera, etc. are just as liable to pick up and compile > a package whose author built in a booby-trap. IMHO, Red Hat, Slackware, Irix, SunOS, Solaris, HPUX are NOT AS LIKELY to INSTALL a booby-trapped package. Since extraction, compilation, and testing are nominally done by an unprivelaged user (e.g. tool.bin) before privelages are granted, a booby-trap has to be clever enough to pass the fitness of purpose testing done by the tool manager. Users, groups, and permissions are used like doors. They separate the bearer bonds (behind the safe door) from the silverware (in the fancy chest) and the phone book (lying on the counter). They separate the food (kitchen) from the pesticides (garage). Valuables like pap-secrets are protected behind superuser privelages. Good stuff (like internet access) may be protected by user privelages. The home page may be unprotected. > We are working on this problem > by establishing a standard for authors to use when signing their software, > and we will work to get authors into the PGP web of trust through our > certification authority or other means (like having a local Debian developer > check them out) so that we can trace software all the way back to the > original author. Author traceability is good, but a central certification authority implies either a substantial barrier to entry (the cost of certification process reliable enough for valuables), or a risk of forgery too high to protect valuables. Author certificates are like badges. Without doors (or with everything from the company advertising calendars to the payroll cash in one room) they are useless. Thank you, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: RPM
Debians, I am a un*x guru, but a Debian newbie. I apologize in advance if the following three questions indicate avoidable ignorance of the proper usage of dpkg. I recently installed dpkg and dpkg-dev 1.4 only to find that it did not remove the obsolete files of dpkg and dpkg-dev 1.2 (e.g. /usr/doc/dpkg/*.txt). A. How can one install debian packages without giving superuser privelages to the person who assembled the package? B. How can one cleanly remove a debian package? C. How can one cleanly remove a debian package that failed to install? I think the answers to these questions are serious enough to decide whether Debian linux will grow or die. > Les Mikesell [EMAIL PROTECTED] Wrote: > What we really need is a way for the installer to set up and maintain > a policy file that establishes the filesystem layout and where > various programs are installed. I don't see how being trapped > into forever using the layout philosophy from some distribution > is a strength for free software. > I do realize that this would be an enormous job for existing packages > but it seems like it could be done for new work. --- My understanding of traditional un*x package management. --- A. How can one install packages without giving superuser privelages to the person who assembled the package? 1. Superuser creates a directory /usr/packages/ and gives ownership to tool.bin. 2. tool (an unprivelaged user) extracts the tar file into directories [bin, lib, etc] under /usr/packages/. 3. tool builds, compiles, configures, tests, etc. the package under /usr/packages/. The ordinary permission system prevents tool (an unprivelage user) from unexpectedly interfering with any other package. 4. (After satisfactory testing) Superuser symbolically links (or copies) the necessary files to where they are avialable to the community. B. How can one cleanly remove a package? 1. find(1) and remove all symbolic links to /usr/packages//... 2. sudo -rf /usr/packages/. C. How can one cleanly remove a package that failed to install? 1. find(1) and remove all symbolic links to /usr/packages//... [1. is seldom (never?) necessary, since they won't be generated until the package installs correctly.] 2. sudo -rf /usr/packages/. [Most system administrators I know used personal scripts to implement a variation of the above. opt_depot is a set of scripts from Denver University(?) that implement the above.] [I personally add a directory /usr/packages//original in which I put the original tar file, its license, description, and a journal of installation, configuration, and maintentance activity.] --- My understanding of traditional un*x package management. --- --- My understanding of Windows and Windows95 answer to the above questions --- A. How can one install Microsoft or other packages without giving superuser privelages to the person who assembled the package? You can't. The package assemblers know everything. Any problem is your fault for having something they didn't know about on your system, such as a package supplied by a competitor, or another product that depends on a different version of a library. B. How can one cleanly remove a package? You can't. The package assemblers provide "uninstall" which will tell you that it removed everything and destroy all traceability of the files that it failed to remove, but still occupy space. C. How can one cleanly remove a package that failed to install? You can't. The package assemblers know everything. Any problem is your fault for having something they didn't know about on your system, such as a package supplied by a competitor, or the results of a past installation failure. --- My understanding of Windows and Windows95 answer to the above questions --- --- My understanding of the consequences of Windows and Windows95 answers --- A. Installation of any package risks the destruction, disabling, or destabilizing of every currently installed package. [This is one source of the Microsoft reputation for products that mysteriously stop working.] B. With time, the disk accumulates cruft whose origin and purpose is unknown. The consequences of removal are likewise unknown, and seldom risked. C. Every upgrade or installation carries the risk that the entire system will have to be reinstalled from scratch. [This largely eliminates any software not received from a single source. In other words, this largely eliminates free software.] --- My understanding of the consequences of Windows and Windows95 answers --- -- Robert Meier Microsoft has a software group colloquially know as the "Wreck a Nice Beach" group. C
Re: gcc stuff
Seth,
Based on your question, I am going to guess that you are not
familiar with C, but are familiar with an object-oriented language (e.g. lisp)
or a stream language (basic). If you will be using C or C++, I recommend
reading a copy of
The C Programming Language by Kerninghan and Ritchie
ISBN 0-13-110362-8
The C++ Programming Language by Stroustrop
???
C++ IOStreams Handbook by Steve Teale
ISBN 0-201-59641-5
C has no intrinsic i/o, but depends on functions provided in
standard libraries to read files. These books will provide the concepts
necessary.
In the meantime, I enclose an example program in C and the same
program in C++.
to use copy.c
1. Copy the program below into files named copy.c.
2. Type "gcc -o copy copy.c" to compile the program.
3. Type "cp copy.c copy.in" to create sample input.
4. Type "./copy" to execute the program.
5. Type "cat copy.out" to view the result.
to use copy.c
copy.c
/*
* Project: C Programming
* File:Trivial file example
* Copy a file from input to output
* Author: Dr. Robert J. Meier
* History: 95-10-24 -rjm- file creation
*/
/* io functions defined in the standard library are declared in stdio.h */
#include /* FILE */
/* Nominal c programs are executed by evaluating the special function, main */
main() {
/* Create a data structure for the input and output file */
FILE *in = fopen("copy.in", "r");
FILE *out = fopen("copy.out", "w");
/* Report an error if the files cannot be openned */
if (!in) {
perror("Unable to read copy.in");
} else if (!out) {
perror("Unable to write copy.out");
} else {
/* If everything is ready copy each character in a loop until EOF */
int c;
for (; EOF != (c = fgetc(in)); fputc(c, out));
}
/* Reclaim memory space from the file data structures */
fclose(in);
fclose(out);
return 0;
}
copy.c
to use copy.C
1. Copy the program below into files named copy.C.
2. Type "g++ -o copy copy.C" to compile the program.
3. Type "cp copy.C copy.in" to create sample input.
4. Type "./copy" to execute the program.
5. Type "cat copy.out" to view the result.
to use copy.C
copy.C
// Project: C++ Programming
// File:Trivial file example
// Copy a file from input to output
// Author: Dr. Robert J. Meier
// History: 95-10-24 -rjm- file creation
// 97-03-24 -rjm- converted to c++
// io classes defined in the standard library are declared in streamio.h
#include// ifstream
// Nominal c++ programs are executed by evaluating the special function, main
main() {
// Open the input and output streams
ifstream in("copy.in");
ofstream out("copy.out");
// Report an error if the files cannot be openned
if (!in) {
cerr << "Unable to read copy.in" << endl;
} else if (!out) {
cerr << "Unable to read copy.out" << endl;
} else {
// If everything is ready copy each character in a loop until EOF
for (char c; c = in.get(), !in.eof(); out << c);
}
return 0;
}
copy.C
With life on loan from God,
--
Robert Meier
FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED]
Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: How to uninitialise partition?
Nick, Having had to recover disks in in the past (on Apples, Macs, IBMS, and SGIs) I offer the folowing hint. If linux fdisk is able to recover the disk, great! If not, and you must extract the files from the disk dump (from dd?), I have found the inode-reading features (-i -C -c -k) of cpio (and afio) very useful. See the manuals for details. If you are not already familiar with the inode system used by *nix, please mail me direct and I can provide explanation or point you at some textbooks that will probably give you more detail than you would ever want. You are not the first (or the last) who has munged a disk without a backup. You are far ahead of most, in that you stopped before further damage was done. Good luck, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: Why is my /home g+s staff? [Was Re: dpkg and shadow]
> Here's a quick question: Why is my /home g+s staff? > drwxrwsr-x 4 root staff1024 Mar 19 23:05 /home/ Your $HOME directory is probably /home/. /home is the parent directory for the home directories of all the users (real and system) on your machine. Normally only the staff can add or delete users, or change access to users' directories. http://www.pathname.com/fhs/1.2/fsstnd-toc.html recommends a good division of the address space and authority. For decades, I have added drwxr.xr.x 4 tool bin... /usr/packages I place original distributions in /usr/packages/ along with their licensing info, installation notes, etc. The package is actually installed here by an unprivelaged user (tool) and then symbolic links connect it to /usr/local/... By this method, one can install packages easily, and uninstall cleanly. Removal of the directory, and all symbolic links into it, will entirely purge the package except for a handful of configuration files that I always list in the installation notes. For the last year, I've been using a package called opt_depot that is a set of shell scripts to automate the above. I plan to put opt_depot on my new Debian system soon to handle packages that are not yet debianized (e.g. kermit, Mosaic, calc-el) and to provide clean removal capability. Hope this helps, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: mgetty
Jason, > Is anyone using mgetty?? Did anyone have problems getting it setup?? I have used for the past year with Slackware 3.0 (linux 1.2.13). I had no trouble using it for fax reception or transmission, and remote login. I have not yet started using it under Debian. For security purposes, I changed /dev/ttyS1 ownership to uucp.uucp and ran the fax daemon as uucp. I did have some trouble configuring it to work with kermit. It frequently detected kermit use as a failed fax reception, and kindly reset the modem. I worked around this by adding a run level to /etc/inittab which disabled mgetty and allowed me to use kermit. An inelegant kludge I admit. A better suggestion from this list would be appreciated. Reporting -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: dpkg list of all packages [Was Re: dpkg and shadow]
> Is there a way to make dpkg give me a list (output) of ALL the packages > (installed and not installed) along with their descriptions? Even just > the short one-liner descriptions? ... I want to search ... by a keyword ... dpkg -l This produces a list of selected packages (installed, half-installed, removed, etc.). Purged packages are not included. I have used this with find(1), dpkg --search, and dpkg -s to obtain any information that I wanted with regards to any downloaded packages. grep /var/lib/dpkg/available This searches the dpkg database cache regarding packages that may be downloaded. See "dpkg --help" for further details. -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: write-protected filesystems [Was Re: dpkg and shadow]
> [When I mount] an Iomega Zip drive ... occasionally[,] I get the message: > sda: Write Protect is off >sda: sda4 > ... Is there a command out there somewhere to write protect/unprotect > disks? I am not familiar with Iomega drives but strongly suspect that the message is simply a warning that the write-protect tab is in the writable position. *nix distinguishes between the drive (a device) and the filesystem, (its data). You can write-protect the drive (disallow control) by using chmod a-w on its node /dev/sda?. This is probably not what you want since reading a device that you can't control is usually not useful. mount -r (see "man 8 mount for further details) will mount the filesystem (data) read-only. All the standard i/o drivers (see man 4 ioctl and man 3 errno for further details) will return -1 and set errno to EROFS. Setting the write-protect tab on the media will normally have the same effect. Hope this helps -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Experience approaching manufacturers for protocols
Debians, I am interested in writing a Linux interface for the PS6500 Personal Data Organizer from Texas Instruments. Before contacting TI, I would appreciate any experience or suggestions regarding approaching manufacturers for protocols. Is there an TI employee on this list who would like to be involved? Does anyone have any experience getting protocols from TI? Does anyone have any experience approaching any manufacturers for protocols? What features of such an interface would interest people? -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: Security Issue
Matthew, > I'm not sure if this is normal, but it seems that any file owned by > someone else and in one of my directories can be deleted by me ... > I also can rename the file, but I can't alter the file. This holds true > even if the file is owned by root. > > Is this normal ? Yes. Permissions are a frequent stumbling block for new unix users. The concept is very simple and powerful. > If so, what things can I do to someone elses file thats in one of my > directories , just delete or rename the file ? Everything in unix is a file. I'll say that again. Everything in unix is a file. A directory is a file. If you will agree not to flame me for hyperbole, it may simplify matters to think of the directory as the card catalog in a library. You can see (read) and deface (rename) the card entry even if the rare volume that it references is under lock and key. If you remove the card, that volume is inaccessible (deleted) to anyone using the card catalog, but note that the book may still be accessible through another card catalog. Deleting the file in one directory does not actually delete the file if there are hard links to it from other directories. The mode (permissions) of a file are nominally handled as triples (r, w, x). r - permission to see the contents of the file contents of the directory (index file) input of the device w - permission to alter the contents of the file contents of the directory (index file) output of the device x - permission to execute the contents of the file the search of the directory (index file) Each file has an assigned user and group. The mode has three triples that apply respectively to the user, group, and others. For further information see "man chmod". If you have write permission on the directory, you have the ability to alter its contents (the file index) including deletion, and renaming. You even have this permission if you don't have read or search permission! > As root, what if I want to keep a file in someones directory without them > deleteing it ? As I see it now, that can't be done ?!?!? Root can delegate authority over a directory by granting write permission to its user. This is the customary permission for home directories. So long as that authority has been delegated, root cannot prevent a user from removing any file including those owned by root. If root needs to keep a file in a directory despite all other efforts, then root must own and deny anyone else write access to the directory and all its parents. It is customary that root own only critical files (/etc/passwd, /etc/ppp/pap-secrets, etc.) and directories (/etc, /bin). Most of the address space is delegated to officers (operator, uucp, mail, etc.) and groups (staff, adm, www, etc.) determined by convention and experience. This separation of authority makes unix much easier to manage for multi-user installations. It also provides a check system that limits the spread of either accidental or malicious interference (viruses). Most professional system administrators that I know usually cringe at the thought of unnecessary use of root and its ability to destroy everything with a few mistaken keystrokes. sudo, doas, and well-chosen group assignments provide a safety net. Hope this helps, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: Upgrading from slackware to debian
Stig, > I wonder if it is possible to upgrade from slackware to debian, and > keep most of my system. Is there an easy way to do this, or should I > just fdisk the thing and start all over? I recently upgraded from slackware 3.0 (linux 1.2.13) to Debian (linux 2.0.6) and still have both resident on my host. It is therefore possible. Following the Debian instructions, I used fdisk and started all over in a fresh partition. This was not easy. In hindsight, I believe that it was safe but not necessary to start fresh. I believe that it can be done much more simply using the slackware environment. If you or others are interested further, please email me direct, so we don't use list bandwidth. Trying to help, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
TI PS6500/6155/6555 organizer and Linux
Debians, I just purchased a Texas Instruments personal organizer (Model 6500) with kit (6555) for bidirectional serial communication (6155) to a PC. Does anyone have experience or recommendations with this model? Does anyone know of or have software to link this to Linux? Is the protocol available? Thank you for your support, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: My screen gets messed up :-(
Leander Berwers, > After I do a 'more (binary file)' often my command prompt and everything I > type is unreadable. This has probably something to do with ANSI. How can I > correct this 'problem'? Don't more (or less) binary files. It's an "ASCII thing". The American Standard Code for Information Interchange divided the space of 256 8-bit character into categories and assigned meanings to each characters. The general concept is called code space multiplexing and divides one communication channel (the serial cable) into many (human<->human, modem<->modem, host<->host, etc.). The codes from 0x20-0x7e make up the "printable ascii" set and are designated for human endusers. The codes from 0x00-0x1f, and 0x7f were assigned specific meanings for modem<->modem and host<->host communication (go to the next line, end of transmission, slow down transmission, etc.) The codes from 0x80-0xff were designated for application use, with some recommended assignments. When you more or less a binary file, the terminal and host interpret the non-printable characters (0x00-0x1f, and 0x7f-0xff). If done intentionally, this adds boldface, tables, invisible fields, animation, line drawing, and other features to available presentation formats. A random binary file is unlikely to have desirable results. It is similar to letting the mythological monkey make your ATM deposit. One way to view the contents of a binary file is to use od -vb . See "man od" for more information. Several editors such as emacs have a hex-mode that allows you to work conveniently with binary files. Welcome to linux, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
Re: mt eof + Travan = Operation not permitted
Jens, Ken, and any other interested parties Thank you for your able assistance with the mt(8). The secret was the need to use mt erase on a tape. This works for linux 2.06, but does not work for linux 1.2.13 (Operation not permitted). I understand that patch 56(?) addressed this problem for linux 1.2.13. If you have time, I am curious. What is the action performed by erase that makes multiple file storage possible? Why is only the first file accessible if erase is not performed? I understood that tape operations recorded nominally-fixed-length blocks of bits on the tape with eof marks in between, and an eot at the end. Tar by default writes several file blocks followed by two eof marks. ...++..++.++* ^^^ ^^ ^file blocks ^^ ^^ ^^ 6 eof marks (2/file by convention) ^ eot mark What is the difference between a pre-formatted (QIC02?) tape and a tape after mt erase? Curious, -- Robert Meier "Albert Einstein in 1906, Set out a theory, the universe to fix. No longer apart, space and time came together, Their wrinkles account for the fall of a feather." Physicists by Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
mt eof + Travan = Operation not permitted
Debians, Thank you in advance for any help. I recently upgraded my Slackware 3.0 (linux 1.2.13) system to Debian 1.1 (linux 2.0.6) via InfoMagic's Dec 96 6 cd set, and was disappointed that mt(1) limitations were not lifted. Summary:How can I read and write multiple archives on a floppy tape? Is anyone writing multiple archives on a Travan floppy tape? Problem:mt -f /dev/nrft0 eof reports "Operation not permitted" I understood that this was a limitation of linux 1.2.13 but not of linux 2.0.6. tar -cvf /dev/nrft0 . tar -cvf /dev/nrft0 ./.deleted writes the first file (archive of .) does not write the second file (archive of ./.deleted) 400MB tapes have ~10%/400MB read error rate. I suspect that the Travan T-1000 motor is only marginally capable of handling the moment of inertia of 400MB tapes. Workaround: (for past year with Slackware 3.0) Place only one archive on each tape. Use /dev/rft0. Use only 200MB tapes. Config: Dell Dimension XPS-133c Travan T-1000 400MB internal floppy tape controller Debian 1.1 linux 2.0.6 Any help appreciated, -- Robert Meier "Democracy is the most inefficient form of government, except for all the others." -- Winston Churchill FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
new binary installation + package trouble => rebuild the kernel
Debians, I would like to suggest to those having trouble installing or using packages in a new binary installation that they consider spending ~6 hours to rebuild and reinstall the kernel. I have recently upgraded my system from Slackware 3.0 (linux 1.2.13) to Debian 1.1 (linux 2.0.6) via the Dec 96 InfoMagic 6 cd set. Of ~200 packages selected, ~10 would not install, 2 refused removal, ~20 were tested, and ~5 worked. Gratefully watching the help provided to other reporting similar problems, I was able to get ~5 more packages working. Last year, after installing Slackware, I wasted ~3mo struggling with printer and other problems. I resisted recompiling the kernel because I remembered this as a 3d-2wk operation from Vax 11/780 days. After I finally spent 2 hours configuring and 30min :) compiling and installing the new kernel, my undiagnosable printer (and most other) problems disappeared. This year, I only wasted two weeks before recompiling the Debian kernel. After 5 hours reading the excellent help provided by xconfig, it took only 30 min to compile and 30 min to install the new kernel. Since replacing the binaries supplied by InfoMagic with the source supplied by InfoMagic and compiled locally, 25 of 26 tested packages have worked with little or no additional effort. Hoping to contribute soon, -- Robert Meier "To bend a beam of starlight, Relativity mechanics say, Watch the sky turn every night, Three billion miles away." - Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
X installation problems
Debians, Thank you in advance from a newbie. Any help in installing X would be appreciated, especially regarding o booting to a login screen o configuring the mouse for left-handed use I have been trying to install Debian 1.2 from theInfoMagick Dec 1996 6cd set using dpkg on aDell Dimension XPS133c withNumber Nine FX Motion 771 graphics accelerator and Trinitron D1526TX-HS Ultrascan monitor and PS/2 3-button mouse dpkg did not complete installation the first time, and would not allow reinstallation, configuration, removal, or purging because it was "already installed" or "in a bad state". I finally used dpkg --force-dependencies to purge xbase and reinstall it. After running xf86config, I could run startx to begin an X session but the screen was unusable. After copying (and editting) XF86Config from my slackware installation, I was able to run startx and use the windows. After hand-editting /etc/inittab, and /etc/rc4.d/S99xdm, xdm is started as a deamon at boot. Unfortunately, though xdm remains running (as seen by ps -ax), no login screen is presented. I have read /usr/doc/HOWTO/HOWTO-X11R6 and man [xdm xinit startx ...] but have been unable to find any way to cause or inhibit presentation of a login screen when xdm starts. My understanding is that when xdm starts, it runs the setup script and then presents a login screen to all hosts that access it. What configuration do I need so that debian boots to a login screen? What configuration (change to XF86Config) is needed to reverse the button assignment (left-handed instead of right-handed use)? Ideally, I would like the following run level assignments 0 - shutdown 1 - single user 4 - multi user (no xdm daemon) 5 - multi user (xdm login screen) 6 - restart Newbie, -- Robert Meier FANUC Robotics North America, Inc. Internet: [EMAIL PROTECTED] Voice: 1-810-377-7469 Fax: 1-810-377-7363
