Changing Users in a script
I have a system with several different users and would like to use cron to run
this script as root:
#!/bin/bash
for user in `ls /home/`; do
# echo Path: $user
if [ ${user:0:1} != 0 ]; then
path=/home/$user/Backup
if [ -e $path ]; then
echo Calling backup for user: $user
sudo -u $user /usr/local/bin/user-backup
fi
fi
done
The idea is that instead of adding a backup script every time I add a user,
this script will go through the /home directories and skip any that start with
a 0 (a program I'm using creates some directories there, but starts their names
with a 0) and automatically call the generic backup script for that user.
The problem is sudo can't be run without a tty, so I can run it myself, but it
won't run from a script.
I want the backup script to run under each user's name to match the user on the
backup system.
Are there other ways to do this with an all-in-one approach? Either for a
script run as root to run scripts with the id of the users or some generic way
to tell cron to run a script once for each user that meets certain conditions?
I prefer the all-in-one solutions, since when I add a user, I'm adding it to
their system, to this backup NAS, and to an offsite backup NAS, and even though
I use notes, it's easy to forget having to do extra things when adding a user.
So I'd really prefer a solution that handles all of them at once.
Any other way I can do this?
Thanks!
Hal
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/6f014e0d-554d-449c-8a28-0e989a86d...@halblog.com
Re: Changing Users in a script
On Aug 15, 2011, at 2:05 PM, Bob Proulx wrote:
Hal Vaughan wrote:
I have a system with several different users and would like to use
cron to run this script as root:
#!/bin/bash
for user in `ls /home/`; do
#echo Path: $user
if [ ${user:0:1} != 0 ]; then
path=/home/$user/Backup
if [ -e $path ]; then
echo Calling backup for user: $user
sudo -u $user /usr/local/bin/user-backup
fi
fi
done
Personally if I were writing this then if the script is running as
root then instead of using 'sudo' here I would use 'su' instead.
su -c /usr/local/bin/user-backup $user
Mostly because su is more traditional and more core than sudo and
just seems like the better fit for the job. But it is a matter of
taste here.
Senile moment -- I keep forgetting that I can use su for different users, too.
Thanks for the reminder!
The idea is that instead of adding a backup script every time I add
a user, this script will go through the /home directories and skip
any that start with a 0 (a program I'm using creates some
directories there, but starts their names with a 0) and
automatically call the generic backup script for that user.
Seems reasonable so far. Also you should skip directories called
lost+found in the case that /home happens to be a mount point on a
filesystem such as ext[23] that uses lost+found.
Thanks for the reminder!
The problem is sudo can't be run without a tty, so I can run it
myself, but it won't run from a script.
Using 'su' would solve that problem.
Any other way I can do this?
There are distinct advantages to a backup push system. Not proposing
that you change away from it. But I tend to pull backups from /home
to the backup server. This means that whatever is in /home comes over
whether it is associated with a user's home directory or not. All I
manage is machines. Not machines and users.
How is it that a pull system would get more? I didn't know there'd be files
associated with a user that are outside of their home directory.
Hal
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/f25320a2-3441-4188-9071-c1218dc59...@halblog.com
Re: Changing Users in a script
On Aug 15, 2011, at 2:15 PM, Bob Proulx wrote: Bob Proulx wrote: Hal Vaughan wrote: The problem is sudo can't be run without a tty, so I can run it myself, but it won't run from a script. Using 'su' would solve that problem. BTW... I assume that is because you have tty-tickets turned on for sudo? In which case you could avoid it with sudo too by turning off tty-tickets for this use case. Not dealing with that. For my two NAS units, I really want to do as little as possible. Since I need the script anyway, I'd rather just use that only. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4e895b5e-1be9-464d-b126-9fdc51617...@halblog.com
Rsync -- Different Outputs on No Transfer
I'm using rsync on normal Debian (6.x), on two embedded systems that run what look like Debian variations (DNS-321 by D-Link and Stora by Netgear) and on OS X. On Debian, whenever I run rsync (rsync --delete -rlptv -e ssh /my/path/ myname@mybackup:Backup/, if there are no files to transfer, rsync prints a long list of directories and I get a high enough count for the bytes transferred that it's well over just the text of the list of directories. With other systems, I often get something more like this (from the same command with the same flags): [admin@server:~]$ rsync --delete -rlptv -e ssh /ServerShare/data/ myname@backup:Backup/data/ Stora version 10.0.x building file list ... done sent 415903 bytes received 20 bytes 12797.63 bytes/sec total size is 5840010926 speedup is 14041.09 I understand the Stora version 10.0.x is from the Stora when I essentially log in via ssh for this backup. I also notice this transfers over 400k bytes with no data sent to the backup system. Is that all checksums and filenames? All these backups are going to the same system, so the difference in whether I get a simple output or a long listing of directories scanned for backup files would seem to be due to something on the sending system. I use the -v flag whenever in case I need to debug later. Any idea why, on Debian, I always get a long directory listing and don't get it on some other systems? It's not a must fix but when I'm scanning output files, obviously it's a LOT easier to verify everything went smoothly if I get a quick and simple output than if I have to scan a long list of directories. It'd be nice to simplify it so I can tell at a glance when things went well. Any suggestions on what could cause the difference? Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/a689e2d2-cfab-4e00-8161-f31226417...@halblog.com
Re: Rsync -- Different Outputs on No Transfer
On Aug 14, 2011, at 10:08 PM, Ivan Shmakov wrote: Hal Vaughan h...@halblog.com writes: […] It's not a must fix but when I'm scanning output files, obviously it's a LOT easier to verify everything went smoothly if I get a quick and simple output than if I have to scan a long list of directories. It'd be nice to simplify it so I can tell at a glance when things went well. Any suggestions on what could cause the difference? My guess is that using -O along with -t may reduce the number of directories in the -v list. Thanks. I tried with -O and without it, along with -t and no -t (in other words all four combinations of those two), but I still get the directory listing. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e93639de-0baf-4364-9646-fe5a7c616...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 25, 2011, at 6:15 AM, Liam O'Toole wrote: On 2011-04-25, Hal Vaughan h...@halblog.com wrote: --- SNIP --- A couple points here. I checked the source code for the Dynamic Tree Demo that is used as a JNLP demo. Now I will admit I haven't been using Java for a couple years and have some catching up to do (I think I was using Java 5, but know I've totally missed Java 6), but when I went over the source code for that demo, there was no main() method in the source. Or, more correctly, there was, but it was commented out. That, right there, is an indication that if you're writing an app in Java to be launched normally, there will be different preparations than if you're using JNLP. (I haven't used JNLP yet, but I did write some applets and even set up a couple Java apps to be both apps and applets.) I can only think that the main() method was commented out later on. A Java application deployed using JNLP has its entry point in the main method, just like any other Java application. See the third bullet point in the development tutorial[1]. OKay, I see that in link you provided. It was in another class - I had looked at it several times but didn't see it. (I always give a line break between methods in a class and anything else and they didn't, due to a learning disability with symbol recognition, so I missed it until I used grep.) I even looked at their more in depth tutorial on deploying a JNLP project, but didn't find any way to launch one, other than through a browser. (Which I don't want to do -- then I have to figure out the default browser from within a Java class on Windows, Linux, and OS X and launch it.) You don't need to do that. You simply need the relevant file association on the client. On a Debian system you have: $ grep jnlp /usr/share/applications/mimeinfo.cache application/x-java-jnlp-file=sun-java6-javaws.desktop Okay, that works for me. One thing I said earlier was that I figure, at this point, I can count on Java being installed on the client's desktop computer. (I'll fill you in on that at the end of the email, since there's a few items I've mentioned already.) I also just tested it on Windows (I only have up through XP and use it in a VM on my iMac) and javaws is in the path. I don't know where the file associations are stored in Windows (probably in the registry), but in my case I think it's a relatively safe gamble that if Java is installed, javaws will be the default to handle a jnlp file. When the .jnlp file is accessed over HTTP using a web browser, the browser passes the file to the local javaws executable. Here the browser is simply arranging the transport. Equally, the user could access the .jnlp file using Windows Explorer or Nautilus, etc. You can even go directly to the file from the command line: javaws http://download.oracle.com/docs/cd/E17746_01/tutorialJWS/deployment/webstart/ex6/webstart_ComponentArch_DynamicTreeDemo/dynamictree-webstart.jnlp I just tested that on Linux, OS X, and Windows. It works on all three of them. I had some trouble on Linux, since it's on a VM and doesn't have X and wasn't set to forward, but the fact that I got a Java error message tells me Java (or javaws) was attempting to run the program and did make it to the class file. (There is an understandable but unfortunate tendency to confuse Java Web Start with applet technology. An applet runs in isolation within a web browser (at least we hope it does). In the case of JWS, once the resources are delivered to the client then the browser or file manager is out of the picture.) And that is a problem I was having. I had never used JWS before and thought it was related to an applet. You need some sort of xml file defining the app to launch and while java does not (want to) understand the smb:// protocol we are stuck. I can indeed play JNLP based java applications that are hosted online but I cannot run the java application I want to run while is on samba share. At least not directly ;-( Try using a UNC path in the JNLP descriptor, e.g., file:server/share/app.jar. Windows clients will understand that. This works on Windows, with SMB, but that brings us back to the original point: Accessing a file on a network share without mounting the share. On Windows, since SMB is built in, you can easily do that. On Linux, you can't. You can use smbclient and copy the file over to your system if it's on an SMB share, but nobody yet has provided a way to access a file on a network share from Linux without either mounting the share or copying the file over to the local computer first. Yes, that is a problem. There's more to the issue than that -- again, see below, so those that saw the first email don't have to read it all. I'm looking at Java RMI, but that involves certificates and so on and I'm still not sure if I will easily be able to put that much
Re: Using Files Without Mounting A Share From Another System
On Apr 25, 2011, at 10:21 AM, Camaleón wrote: On Sun, 24 Apr 2011 20:33:45 +, Liam O'Toole wrote: On 2011-04-24, Camaleón wrote: (...) Once the user trusts the application, the local JVM will load and run the JAR file exactly as if it had originally resided on the local hard drive. Yes, but not all java applications are ready to be used with that method. If you can tell me a sample case, I will test :-) --- SNIP --- I cannot provide you with a sample case to prove a negative! Instead I invite you to download the sample .jnlp file in the link I provided earlier and modify it to suit your LAN. Okay, I just have tried to create a JNLP file to launch the JAR file that is stored on a samba share but I'm still facing the same problem that we are encountering from the beginning: smb:// protocol is not detected ;-( If I place the test.jnlp file locally it can be run (it fails because it looks for a signature file that seems to be missed from the original jar but anyway it recognizes the JAR and tries to launch it), but when I put the same test.jnlp file over a samba share and run the app from a linux client it does not work (java.io.FileNotFoundException: smb:/host/ share/test.jnlp (No such file or directory). Needless to say the same test.jnlp that is located on the network share runs fine when it's launched from a windows client ;-( I think this setup will solve both our problems. But don't try to launch it from a network share, set up Apache (or another web server) and launch it through there. I tried Liam's example from the command line and it works perfectly -- it just means we have to put the program in a directory where Apache can serve it, instead of sharing it through SMB or NFS. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5c22b99f-f222-4756-9b3e-43cd62381...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 25, 2011, at 1:05 PM, Liam O'Toole wrote: On 2011-04-25, Camaleón noela...@gmail.com wrote: On Sun, 24 Apr 2011 20:33:45 +, Liam O'Toole wrote: On 2011-04-24, Camaleón wrote: (...) Once the user trusts the application, the local JVM will load and run the JAR file exactly as if it had originally resided on the local hard drive. Yes, but not all java applications are ready to be used with that method. If you can tell me a sample case, I will test :-) --- SNIP --- I cannot provide you with a sample case to prove a negative! Instead I invite you to download the sample .jnlp file in the link I provided earlier and modify it to suit your LAN. Okay, I just have tried to create a JNLP file to launch the JAR file that is stored on a samba share but I'm still facing the same problem that we are encountering from the beginning: smb:// protocol is not detected ;-( If I place the test.jnlp file locally it can be run (it fails because it looks for a signature file that seems to be missed from the original jar but anyway it recognizes the JAR and tries to launch it), but when I put the same test.jnlp file over a samba share and run the app from a linux client it does not work (java.io.FileNotFoundException: smb:/host/ share/test.jnlp (No such file or directory). Needless to say the same test.jnlp that is located on the network share runs fine when it's launched from a windows client ;-( That's a pity. I suspect that smb:// it is not supported by Java because, like sftp://, it is not an official scheme[1]. I'm almost positive SMB is not supported by Java and it's Windows/DOS that supports it when you're testing. I found I can list a directory on an unmounted SMB share from DOS with dir \\server\\dirctory. Also, if you go to Samba's website, somewhere on there is a CIFS jar that provides SMB/CIFS for Java. If it were already supported, I doubt they'd be doing that. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ad3196cf-0511-4c39-b0dc-25fea001c...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 25, 2011, at 2:53 PM, Liam O'Toole wrote: On 2011-04-25, Hal Vaughan h...@halblog.com wrote: --- SNIP --- On Apr 25, 2011, at 6:15 AM, Liam O'Toole wrote: --- SNIP --- Are you in a position to put a web server on the machine which offers the network share? That would allow you to serve up a .jnlp file and its associated JAR file in a simple, standard and platform-neutral way. Yes, I can put apache on the server without a problem. Considering it's an embedded system and only has 500 MB and 500 Mhz, I was hoping I wouldn't need it, but I can use it. Consider using a lightweight alternative such as thttpd. Thanks -- sometimes you get so used to one package as a standard you forget about the others! I'll stop evangelising now :-) No, don't stop. You've actually solved my problem! I know I can find this through experimentation and more research, but just in case you know off the top of your head, will the Java program being run through JNLP have a sandbox like an applet does? (I know they're different, but will JNLP programs have a sandbox?) And do you know if I can access $PATH (or the Windows equivalent) in the JNLP program so I could call another executable on that computer if I needed to? Once the application is signed, and is accepted by the user, it will be granted unrestricted access. I have to look into this. I remember reading up on it once before. Since my app is on a server on their LAN, I'm trying to make sure they're not going to an outside website for anything and dealing with certification would require that. For accessing environment variables and calling executables, see the java.lang.System and java.lang.Runtime classes, respectively. That approach is frowned upon, however, because it implies writing platform-dependent code and is not considered pure Java. But if it gets the job done ... The other way the client will need to access the server is to run Abiword on the server, with X forwarding set up. On Windows, that means running Xming at \\myserver\\directory\\xming.exe and using Portable PuTTY, which will be at \\myserver\\directory\\putty.exe, but both OS X and Linux will use ssh with X forwarding enabled, so I'm already having to do platform specific stuff, just not much of it. The client has to be able to run Abiword on the server with X-forwarding working and unless I get Java ssh classes AND an x server (that doesn't have any lag), I think I'm going to have to handle it with platform specific code. Other than that, it should all be universal code. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/83f5b785-7700-4258-afe6-0e0bf522f...@halblog.com
Re: Using Files Without Mounting A Share From Another System
Liam, I've read your posts on this thread, but I'm replying in kind of a reverse order -- or skipping the earlier ones if my comments on those make it in here. On Apr 24, 2011, at 4:33 PM, Liam O'Toole wrote: On 2011-04-24, Camaleón noela...@gmail.com wrote: On Sun, 24 Apr 2011 18:08:47 +, Liam O'Toole wrote: On 2011-04-24, Camaleón wrote: Java technology offers a standard way of doing that: Java Web Start. It is possible to deploy a third-party application using a JNLP file[1]. Yes but no, that's a completely different approach (AFAIK, the application has to be prepared/packaged to use it and most of the java apps out there are not intended for that purpose, meaning they are not going to be launched/run online). It is simply a means of lauching an application over the network. Most specifically over http protocol, I'd say. It is most commonly used over HTTP, but other protocols are available (see below). Whether the application was intended to be launched in that way is irrelevant. Nope, it isn't :-) Oh yes it is. Try it. A couple points here. I checked the source code for the Dynamic Tree Demo that is used as a JNLP demo. Now I will admit I haven't been using Java for a couple years and have some catching up to do (I think I was using Java 5, but know I've totally missed Java 6), but when I went over the source code for that demo, there was no main() method in the source. Or, more correctly, there was, but it was commented out. That, right there, is an indication that if you're writing an app in Java to be launched normally, there will be different preparations than if you're using JNLP. (I haven't used JNLP yet, but I did write some applets and even set up a couple Java apps to be both apps and applets.) I even looked at their more in depth tutorial on deploying a JNLP project, but didn't find any way to launch one, other than through a browser. (Which I don't want to do -- then I have to figure out the default browser from within a Java class on Windows, Linux, and OS X and launch it.) You need some sort of xml file defining the app to launch and while java does not (want to) understand the smb:// protocol we are stuck. I can indeed play JNLP based java applications that are hosted online but I cannot run the java application I want to run while is on samba share. At least not directly ;-( Try using a UNC path in the JNLP descriptor, e.g., file:server/share/app.jar. Windows clients will understand that. This works on Windows, with SMB, but that brings us back to the original point: Accessing a file on a network share without mounting the share. On Windows, since SMB is built in, you can easily do that. On Linux, you can't. You can use smbclient and copy the file over to your system if it's on an SMB share, but nobody yet has provided a way to access a file on a network share from Linux without either mounting the share or copying the file over to the local computer first. I'm looking at Java RMI, but that involves certificates and so on and I'm still not sure if I will easily be able to put that much on a client system that is trying to run the program. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9a8b942f-8159-4572-9889-ee36ea2be...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 23, 2011, at 6:44 AM, Camaleón wrote: On Fri, 22 Apr 2011 15:42:56 -0400, Hal Vaughan wrote: On Apr 22, 2011, at 1:56 PM, Camaleón wrote: On Fri, 22 Apr 2011 13:03:00 -0400, Hal Vaughan wrote: (...) I need a way, on Linux, to access files on a network share, which could be SMB or NFS (or something else) without mounting the volume. For example, if I'm on System A and I have an executable on System B, and it's on a network share on System B, is there any way to run that executable without mounting that share as a volume on System A? Hum... I think it could be possible, just ensure that the file in the share has the proper rights (that is, it should be executable by the user). As for java files, you could create a launcher on the desktop pointing to the file: java -jar smb://path/to/jar/file.jar Does Java handle the SMB protocol on its own? I know I can't list a directory that way with ls, even with the Samba client package installed. Good question. Nowadays it should handle smb:// or other network protocol just the same it does with http:// but maybe it has auto-imposed some limitations on linux environments (at least under windows you can launch a java JAR that is stored in a network share) or is just the JAR file has to be prepared to be run over the network (IIRC, there is a jCIFS library to that precisely purpose). I tried, on the command line in Windows XP, dir \\server\directory and it works, but SMB/CIFS is built in to Windows. As far as I can tell, there is no equivalent on Linux unless you use smbclient. SMB is not built into Linux and it seems without smbclient, Linux cannot access an SMB share. I've been searching, and it looks like there's a CIFS class for Java, but it's not built-in to Java and is available at samba.org. I don't have to use SMB, I could have the server use NFS as well, but I can't find anything about reading an NFS share unless it's mounted. As best I can tell, unless you use a program like smbclient, scp, ftp, or rsync, there is no way to read any network share in Linux unless it's mounted. To make a program on a non-local volume available locally, you can either copy it to the local computer or mount the share. At least that's the best I can find out. It's kind of tough to Google this, since searches always turn up tutorials on how to mount a volume. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/3c37ba06-3516-41d7-869e-0b6a5913b...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 23, 2011, at 11:46 AM, Eduardo M KALINOWSKI wrote: On 04/23/2011 12:04 PM, Hal Vaughan wrote: I don't have to use SMB, I could have the server use NFS as well, but I can't find anything about reading an NFS share unless it's mounted. As best I can tell, unless you use a program like smbclient, scp, ftp, or rsync, there is no way to read any network share in Linux unless it's mounted. To make a program on a non-local volume available locally, you can either copy it to the local computer or mount the share. At least that's the best I can find out. It's kind of tough to Google this, since searches always turn up tutorials on how to mount a volume. I don't know how SMB work internally, but I wouldn't be surprised if mounting were just a convenience to ease access to files on the remote computer and to store credentials, but internally each access to a file in under a cifs-mounted path is exactly the same as an access via smbclient or other mount-less programs. Anyway, what's the problem with mounting? I have several reasons for mounting that I'd rather not get into. The short version is that I'm self employed and if a client has had incompetent techs fsck up their LAN in the past and is scared of anything that looks at all different on their LAN, but they also want to pay me $500 - $1,500 a month, it's in my best interest to let them write me that monthly check. (Yes, I've tried educating them, but fear is a strong motivator against learning something that is different from what you are sure is true.) There are a couple other reasons, but that one is enough right there. There is only one way to access files that all programs support: the filesystem. So I think it is not possible to find a general solution to your problem. Particular programs might have extra support, so it is possible to include SMB in some program and then it will not have to mount the remote directory. But that is specific to some program. Unfortunately, I think you're right, but there are a lot of very knowledgeable people on this list, so I was hoping there might be something obscure someone knew about that I could use. This is a cross platform issue, since the executable has to work on Windows, OS X, and Linux. On Windows I can access the SMB share with no difficulty. I'm considering, for Linux and OS X, the idea of just using ssh with X forwarding and running the program on the server and displaying it on the client, but since the server only has 500 MB and 500 Mhz, I'd rather not use any resources on it that I don't have to. I'm also looking into creating a temporary mount point and deleting when it's done, but if there's a crash or something, that temporary mount point would remain, unless I put it in /tmp or something radical like that. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/109e1565-2b5b-44b8-948d-2eb6827e5...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 23, 2011, at 12:00 PM, Camaleón wrote: On Sat, 23 Apr 2011 11:04:20 -0400, Hal Vaughan wrote: On Apr 23, 2011, at 6:44 AM, Camaleón wrote: Does Java handle the SMB protocol on its own? I know I can't list a directory that way with ls, even with the Samba client package installed. Good question. Nowadays it should handle smb:// or other network protocol just the same it does with http:// but maybe it has auto-imposed some limitations on linux environments (at least under windows you can launch a java JAR that is stored in a network share) or is just the JAR file has to be prepared to be run over the network (IIRC, there is a jCIFS library to that precisely purpose). I tried, on the command line in Windows XP, dir \\server\directory and it works, but SMB/CIFS is built in to Windows. As far as I can tell, there is no equivalent on Linux unless you use smbclient. SMB is not built into Linux and it seems without smbclient, Linux cannot access an SMB share. I've been searching, and it looks like there's a CIFS class for Java, but it's not built-in to Java and is available at samba.org. Yep... and it's quite strange. I still don't know if it's a security measure or just a technical barrier that needs the use of another applications to be bypassed. In fact, I can run a .jar file over smb:// using Nautilus but this is what I get: Do you mean you could run the jar or that you tried to? 1/ Jar file is executed but it opens with file roller (archiver utility) 2/ When I try to change file permisson to make it executable I get an error (cannot change permissions to file.jar) Did you try to change permissions from the server itself or from the client? It sounds like you tried to change permissions from the client. All this done over a NTFS volume managed by a Windows host. Maybe a samba share over a linux filesystem (ext3/4/reiser/xfs...) gives different results... dunno :-? I think Linux might give better control over things like permissions. In my case, the server will be running Debian 6.x. Thanks for trying that -- I'm still searching and testing, but I haven't gotten anything to work and I need to create a test jar file I can work with. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4da97253-e3c7-4415-b1a4-84f260d51...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 23, 2011, at 12:20 PM, Eduardo M KALINOWSKI wrote: On 04/23/2011 12:59 PM, Hal Vaughan wrote: I'm also looking into creating a temporary mount point and deleting when it's done, but if there's a crash or something, that temporary mount point would remain, unless I put it in /tmp or something radical like that. Couldn't you copy the file instead (possibly calling smbclient) to the temporary directory? Actually, that's a good point and it would be a lot better than putting a mount point in there. I've just been so keen on the whole don't put it on the client system thing that I didn't think of that. It looks like I'll have to be taking a multi-part approach, using one way for Windows and another for Linux and OS X. And the few people that I deal with that have Linux on a desktop are more aware and don't freak at the idea of having more installed on their computer. And, in my experience, OS X users aren't nearly as tough to deal with in situations like this, either. (It really doesn't make sense -- but then fear often is irrational. I've seen small business owners that freak at the idea of anyone touching their router/firewall to add a static IP address or any port forwarding because their cousin or brother or son fscked up the network when he said he knew what he was doing and they freak at the idea of a small service provider like me installing software on their computer, but then they browse unsafe porn sites and install something off the Internet (saying, But it's on a professional looking website, so they must know what they're doing!) and they think there's no way they're doing anything dangerous.) Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/311b8b0d-0017-47db-8bbb-daaf7edb1...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 23, 2011, at 3:51 PM, Rob Owens wrote: On Fri, Apr 22, 2011 at 01:03:00PM -0400, Hal Vaughan wrote: I now know I can use smbclient to read files on an SMB share without having to mount it, but I need to do more than that. I want to be able to access either Java classes or an executable on a shared volume on a server without having to mount the volume on the local system. (There are a couple reasons for not wanting to mount.) I know on Windows I can list the files on an SMB share on another system and access them using SMB/CIFS by just specifying the volume properly on the command line. I want to do something like that on Linux, but do more than just listing the files or copying them to the local computer. I need a way, on Linux, to access files on a network share, which could be SMB or NFS (or something else) without mounting the volume. For example, if I'm on System A and I have an executable on System B, and it's on a network share on System B, is there any way to run that executable without mounting that share as a volume on System A? Here's a possible workaround. It involves mounting, but as a regular user. I'd use sshfs. The remote server needs to have an ssh server running. Then you can run this: sshfs remoteserver:/some/path localdir Then you can ls localdir, or operate on any of the files there. If you use public key authentication and ssh-add, you can do this without needing to enter a password. I've never used this to access a non-linux machine, but in theory it should work on anything that is running an ssh server. That is mounting, but, as I mentioned in another email, clients that use Linux on a desktop are a lot easier to deal with on these things than Windows users. I still have to test on Windows to be sure that I can actually run a jar from an unmounted SMB share. I can operate on files like that, so I would think I can run a jar that way. sshfs sounds like a good idea and I'll look into it. That would work on both the Mac and Linux. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9d508e39-7be0-4b99-b890-43db96c19...@halblog.com
Using Files Without Mounting A Share From Another System
I now know I can use smbclient to read files on an SMB share without having to mount it, but I need to do more than that. I want to be able to access either Java classes or an executable on a shared volume on a server without having to mount the volume on the local system. (There are a couple reasons for not wanting to mount.) I know on Windows I can list the files on an SMB share on another system and access them using SMB/CIFS by just specifying the volume properly on the command line. I want to do something like that on Linux, but do more than just listing the files or copying them to the local computer. I need a way, on Linux, to access files on a network share, which could be SMB or NFS (or something else) without mounting the volume. For example, if I'm on System A and I have an executable on System B, and it's on a network share on System B, is there any way to run that executable without mounting that share as a volume on System A? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e3f940e5-5e7c-44a6-9dba-36bd272ca...@halblog.com
Re: Using Files Without Mounting A Share From Another System
On Apr 22, 2011, at 1:56 PM, Camaleón wrote: On Fri, 22 Apr 2011 13:03:00 -0400, Hal Vaughan wrote: (...) I need a way, on Linux, to access files on a network share, which could be SMB or NFS (or something else) without mounting the volume. For example, if I'm on System A and I have an executable on System B, and it's on a network share on System B, is there any way to run that executable without mounting that share as a volume on System A? Hum... I think it could be possible, just ensure that the file in the share has the proper rights (that is, it should be executable by the user). As for java files, you could create a launcher on the desktop pointing to the file: java -jar smb://path/to/jar/file.jar Does Java handle the SMB protocol on its own? I know I can't list a directory that way with ls, even with the Samba client package installed. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/11735cf9-4a87-4a52-8d5a-6fd03eeca...@halblog.com
Re: Ubuntu Versions
On Apr 5, 2011, at 7:15 PM, Ron Johnson wrote: On 04/05/2011 05:37 PM, Mark wrote: [snip] upgrading. For all its flaws, one nice thing about Windows is that it has a 10-year (14-year for XP) support cycle, so while there may be service packs, etc., to the end user, the interface is virtually the same for 10 years. I realize that not upgrading/getting more goodies/etc. is not the preference of most people on the list, but for some Debian users it might be. It's an if it ain't broke, don't fix it type thing. You need to mention that to GNOME. Not that they'd list. We're the experts. We know best. That's pretty much why I stopped filing bug reports or issues with almost all FOSS. Too many devs are like that -- too busy being right to listen (and more interested in clearing the bug report than in fixing it). And I say that as someone who retired early because his work as a developer/programmer did what it needed to and made his clients happy and someone who has contributed code to FOSS projects and released some of his own FOSS projects. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/c4d46f7f-90f4-4efa-b55f-02f20561f...@halblog.com
rsyncd.conf - Specifying /home/user for Daemon?
Anything I search for to find this keeps giving me links to backing up your home directory, and I know how to do that. What I'm concerned with is setting the home directory in /etc/rsyncd.conf. In other words, if I type: rsync thisdir me@backupsys::Backup How can I specify in rsyncd.conf that it's using /home/me/Backup? I've tried something like this: [Backup] path = $HOME/Backup/ Comment = Backup service for each individual user read only = no and I can't get that to work. I tried using ~/Backup for the path and using RSYNC_USER_NAME and %u, but that won't work. So how can I specify, in rsyncd.conf that the path leads to a directory in the home directory of the logged in rsync user? Thanks for any help on this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/757e14f4-d705-48e5-b4ce-d03553ec7...@halblog.com
Re: How do I clone Computer A from Computer B?
On Mar 3, 2011, at 12:42 PM, Jason Hsu wrote: Computer A is running minimal Debian with a firewall and servers, including SSH. I can use Computer B to ssh my way into Computer A. How do I use Computer B to clone Computer A? So far, I've only been able to clone Computer A by booting up a live CD on Computer A and running PartImage. I'm not clear if you want to clone A to B, or if you're planning on using A as an image for later installs. In other words, by cloning A, it could be your intent to take the cloned image and install it in other computers along the way. By use Computer B to clone Computer A, it sounds to me like you want to create a cloned image of Computer A that can be stored on Computer B. That's what leads me to wonder if the intent is to create an image of A, and it would seem a most likely use of that image would be to use to create new systems easily and quickly. If that is your purpose, I just went through that in creating an image for embedded systems (specifically a Soekris Net5501). There's one issue nobody's mentioned here: now Debian (and a lot of distros) keeps track of drives and partitions with UUIDs since people are using portable RAM drives now. This effects GRUB2 and /etc/fstab. So if you clone A to an image or to B, be sure to be aware of the issues with UUIDs. This can also create a problem with MAC addresses, too. When using a cloned image moved from one set of hardware to another, I had to update /boot/grub/grub.cfg for the first boot, but also edit /etc/default/grub and make sure, after that first boot, that I ran update-grub and regenerated /boot/grub/grub.cfg (it's also possible to stick the UUID in /boot/grub/grub.cfg if you have a cloned image where you can edit the files). The other problem that is easy to run into if you are, by chance, cloning the image to a computer without a keyboard or monitor, is that unless it has an assigned IP address, you won't know where it is on the LAN. I solved this by writing a simple two-part program that makes it easy to find the new computers that use the cloned image I'm using. I have an image I generated for use on the embedded system I mentioned and it has a Perl script that makes the needed modifications to get the image working on other hardware. It also had the Perl scripts I used to make that system easily locatable on an LAN at http://halblog.com/SqueezeOnSoekris.html. The Perl script that handles updating the image is pretty easy and if you look through it, you'll see all the changes it makes to the cloned image to make it easy to install on a new system without duplicating things like the host name. While this covers more than what you asked, I hope it helps since these are issues that you'll run into these days when you run a cloned system on hardware other than that which the source of the clone runs on. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/c59e482e-c6a6-4f06-8984-9bab8bcde...@halblog.com
Re: Please help
On Mar 4, 2011, at 12:26 AM, Boyd Stephen Smith Jr. wrote: In 20110303230321.8e7ff15c.jhsu802...@jasonhsu.com, Jason Hsu wrote: I may get hate-mail for saying this, but I don't recommend Debian as your first distro. Assuming you have enough RAM (at least 512 MB, preferably 1 GB or more), I recommend Linux Mint. It's known for being user-friendly and accomodating to Windows users. Wrong mailing list. This list is for Debian users, not MS Windows users. Please don't give Microsoft the word windows. If you mean the MS Windows operating system say MS Windows. If you mean the things used in a windowing system (like X11) or those transparent things built into walls, use windows, or Windows is it is the first word of a sentence. I wish it were true, but if it were, then Lindows wouldn't have run into trouble for infringing on the Windows trademark owned by Microsoft. You and I may say this and it may be our opinion, but until it's made clear, legally, that it's not a trademark, it's only our opinion, and we all know there's a big difference between opinion and fact -- no matter how much we want our opinion to be true. Now, if Apple and their app store proves Windows isn't a trademark, then you will be right. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9a3c8437-2e1b-4f25-a3ea-2cd3d3c79...@halblog.com
Re: Please help
On Mar 3, 2011, at 11:06 PM, Ryan Collins wrote: Someone for the love of GOD, help me. I've installedDebian 6.0(squeeze) on my dell inspiron 6000. All I have to show for it is a shell. I've looked all over and read countless entries on ways to load a gnome desktop, but nothing I try works for me like it has others. Im brand spanking new to Linux, but im beginning to at least understand how things need to be typed in order to make the magic happen, but it hasn't happened for me yet. I think I may have installed a server version, but there is a way around that right? What can I do to get a desktop environment installed through my command line? You've already received some good advice, but I'll add a couple points: 1) Server version vs. desktop: That's not a real issue with Linux, especially Debian. On some operating systems it might be due to marketing and so on. In order to sell more and to use a pricing structure that brings in the most income, they might have a home edition and a serve edition. We don't have that on Debian. Instead we use packages. A package can be a font collection, or a program, or graphics, or some other collection of files. If you want OpenOffice, you install that package and it's there. If you want a web server, you install the right packages and you have a web server. If you want a desktop computer, you install the right packages and you have a desktop computer that will work well with a graphical interface and your mouse. But the really cool thing is that installing packages to make the computer a server does not force you to use it ONLY as a server. So even if you've told it to be a server, it can still be a desktop system. Or, if you prefer, you can always restart the install process, since there's nothing tracking how many times you install it or billing you for each install or restricting you to a certain number of installations. 2) Some have suggested using a different distro. I like Debian, but it is designed for certain situations and you may find Linux Mint or some other distros, including Ubuntu, as easier to use when you start out than Debian. I know when I started I used a distro called Mandrake (R.I.P.) that was VERY easy to use and it helped me learn a LOT about Linux. It was a while before I was finally ready to start using Debian (but Debian was also harder to learn in those days). I hope this helps. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/59d0d6d2-4258-4995-baa9-968850e0a...@halblog.com
Re: please help me write a udev rule
On Mar 2, 2011, at 9:08 PM, Yuwen Dai wrote: Dear all, I always run a command like this when I plug a USB disk: echo 1024 /sys/block/sdb/device/max_sectors however, the disk is not always `sdb'. Can I write a udev rule according to the UUID of the disk? Do you mean a specific USB drive you have, or any time a USB drive is plugged in? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2941a649-eb2d-4e7f-986f-e856c2d80...@halblog.com
Re: xinetd UDP vs. TCP Output
Short answer: It's not possible. Long answer: After the research it took me, I'm just too damned lazy to write it up. Just trust me, can't be done. Hal On Feb 24, 2011, at 3:49 AM, Hal Vaughan wrote: I'm using a small program that's started by xinetd. The incoming signal to it would be a broadcast signal, which means it has to be UDP. I wrote two versions of the test program, one in Perl and one as a bash script and both ran into the same problem. They worked fine when I first set them up and set up the service in xinetd as using TCP. Then I changed the service to UDP and made the appropriate changes to my programs. They still logged everything, they still received incoming messages from the other test programs that were communicating to them (either directly or through a broadcast), but they did NOT send any data back. I checked it this with Wireshark. The incoming data showed up, but the data these programs were supposed to send back didn't even go out over the LAN. The programs ran and exited properly, but the output to the network never showed up. While I don't know but so much about networking, I know TCP and UDP sockets are notably different. I can't find anything in the documentation that indicates that for a program using UDP sockets, that it has to use something other than STDIN and STDOUT. I even found sources that say you SHOULD still be using STDIN and STDOUT for programs using UDP through xinetd. I don't consider this just a programming question, since it's the same in both languages. I strongly suspect there's a different way to handle the output that's supposed to go back over the network for UDP (vs. TCP). Any ideas on what might be needed? Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/088447f7-8cad-47ad-ba76-7d3346c7a...@halblog.com -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/b53fd3ac-a847-42eb-8126-43c92b955...@halblog.com
xinetd UDP vs. TCP Output
I'm using a small program that's started by xinetd. The incoming signal to it would be a broadcast signal, which means it has to be UDP. I wrote two versions of the test program, one in Perl and one as a bash script and both ran into the same problem. They worked fine when I first set them up and set up the service in xinetd as using TCP. Then I changed the service to UDP and made the appropriate changes to my programs. They still logged everything, they still received incoming messages from the other test programs that were communicating to them (either directly or through a broadcast), but they did NOT send any data back. I checked it this with Wireshark. The incoming data showed up, but the data these programs were supposed to send back didn't even go out over the LAN. The programs ran and exited properly, but the output to the network never showed up. While I don't know but so much about networking, I know TCP and UDP sockets are notably different. I can't find anything in the documentation that indicates that for a program using UDP sockets, that it has to use something other than STDIN and STDOUT. I even found sources that say you SHOULD still be using STDIN and STDOUT for programs using UDP through xinetd. I don't consider this just a programming question, since it's the same in both languages. I strongly suspect there's a different way to handle the output that's supposed to go back over the network for UDP (vs. TCP). Any ideas on what might be needed? Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/088447f7-8cad-47ad-ba76-7d3346c7a...@halblog.com
Can't Create SymLinks on CF Card
I found this when using tar to unpack an OS image archive. I've been doing this for a while on another CF card and had no problem. But now, when I untar the archive, I keep getting cannot create symlink and operation not permitted. I bypassed tar and tried creating some of the symlinks by hand and got the same error. The CF card is mounted in /mnt/flash. I umounted the drive, cd'ed back to /mnt/flash and ran the very same tar command again and it created the files and symlinks without any problem. So I can create the symlinks to that directory when nothing is mounted, but when I have a CF card mounted there, I can't create symlinks (and have other problems with tar). For example, if the boot directory is there, with the files in it, and vmlinuz is in /mnt/flash/boot, then I do cd /mnt/flash and ln -s boot/vmlinuz ./ I get the same error: Operation not permitted. But if the drive is NOT mounted and I create (from the archive) /mnt/flash/boot, with the same file structure, and do the same two commands to get in /mnt/flash and create the symlink, it works just fine. The only difference is that at one point the files exist on a CF card and the other time they're on the virtual hard drive (this is in a VM in Parallels, the guest OS is Squeeze). I've tried this with another CF card and it works fine. I've even tried it in another mount point and it works. The only time it does NOT work is with this particular CF card. I can unpack files to it and they are there and verified, but I can't create symlinks on the card. Unfortunately, ln doesn't provide any useful information with the verbose mode (other than listing the files it's working on). Does anyone have any idea why I can create and save files on this CF card and cannot make symlinks on it? Thanks for any thoughts on this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2d7e5954-b4cc-430f-8587-a902426e3...@halblog.com
Re: Can't Create SymLinks on CF Card
On Feb 22, 2011, at 4:31 PM, Sven Joachim wrote: On 2011-02-22 21:14 +0100, Hal Vaughan wrote: I found this when using tar to unpack an OS image archive. I've been doing this for a while on another CF card and had no problem. But now, when I untar the archive, I keep getting cannot create symlink and operation not permitted. Looks like the card is formatted with an unsuitable filesystem (e.g. FAT32). Does anyone have any idea why I can create and save files on this CF card and cannot make symlinks on it? Most probably because the filesystem on the card does not support symlinks. You're right on target with that. This is part of a program to install Squeeze on an embedded system and a Compact Flash card wasn't working, so I made a quick trip out to buy a few more. In that time I forgot I had commented out the command in the code to partition the CF card and format the partition. The report I got from sfdisk had the same numbers for the partition size as what I was specifying, so I saw the numbers and assumed it was okay -- without realizing it specified the FS was FAT. So it was my fault -- I forgot my program was not formatting and read the numbers and not the additional data. It's behaving nicely now. Thanks for picking up on something that was so obvious I missed it! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9b2c27ee-3cf8-420d-b606-32b4c0cd0...@halblog.com
Command Line Partitioning
I have a compact flash card I'm experimenting with that goes into an embedded system. In the future I'll be working with more CF cards and I know I won't always know the size of the card ahead of time. I'm working on a Perl program so I can put the CF card in and the program will partition it, copy over an image of Squeeze (copy files, not through dd or anything like that), make the needed changes to GRUB2 (which is a real PITA now that UUIDs are ubiquitous) and anything else that will prepare this CF card for use in the embedded system. I have everything else worked out so it can be done in a batch mode or in a program, but I'm having trouble with partitioning. I've been working with cfdisk, fdisk, and parted: parted: Always leaves only 512 bytes at the start and grub-install doesn't seem to like that. But it can be used in a batch mode. fdisk: Doesn't seem to have a batch/script mode and I'd have to calculate sizes in megabytes from cylinder info cfdisk: Lets me work in whatever units I like (bytes, kbytes, mbytes, and so on) but is curses based and doesn't have batch/script mode. What I would like to do is to be able to use a partitioning program from a batch or script mode. Parted would be good for that, but I can't seem to get it to leave more than 512 bytes at the start (cfdisk left 32k -- why? I don't know). Also, it seems when I get CF disks they have an MS-DOS partition table (grub kept noting that and one part of the whole process didn't seem to like that). How can I either wipe the disk without a mess (tried zeroing it, parted wouldn't recognize it or write to it, but fdisk had no issue with that) or write a new partition table? What type of partition table should I use? (I'm guessing bsd, which was one of the options, but there wasn't a Linux option for types of partition tables.) Thanks for any help on this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4fa96548-a851-4651-bffa-52ca7bd81...@halblog.com
Re: Command Line Partitioning
On Feb 19, 2011, at 12:41 PM, Rob Owens wrote: On Sat, Feb 19, 2011 at 11:49:41AM -0500, Hal Vaughan wrote: fdisk: Doesn't seem to have a batch/script mode and I'd have to calculate sizes in megabytes from cylinder info fdisk does allow you to specify the start cylinder, then something like +1000M for the end of the partition. M meaning megabytes. It may also accept G for gigabytes these days, but I'm not sure -- you'd have to try it yourself. I know I can specify the offset, but there's still the issue of having to calculate numbers to get that 1st cylinder -- and I can't use it from a script. It's interactive. Thanks, though. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/7f1d6133-7f72-41ec-bf84-e9bddb177...@halblog.com
Re: Command Line Partitioning
On Feb 19, 2011, at 4:07 PM, Ron Johnson wrote: On 02/19/2011 02:18 PM, Hal Vaughan wrote: On Feb 19, 2011, at 12:41 PM, Rob Owens wrote: On Sat, Feb 19, 2011 at 11:49:41AM -0500, Hal Vaughan wrote: fdisk: Doesn't seem to have a batch/script mode and I'd have to calculate sizes in megabytes from cylinder info fdisk does allow you to specify the start cylinder, then something like +1000M for the end of the partition. M meaning megabytes. It may also accept G for gigabytes these days, but I'm not sure -- you'd have to try it yourself. I know I can specify the offset, but there's still the issue of having to calculate numbers to get that 1st cylinder -- and I can't use it from a script. It's interactive. If you're always using partition #1, then shouldn't the calculations always be the same, and thus pre-computable? The start, but there are cases where I might want the partition to be larger than in other cases, and I'd still have to figure out the start for the swap partition. On that note -- I noticed that when Squeeze created the swap partition on my original CF card drive, it made it an extended partition and on the test copy, it automatically created a swap after boot that was also an extended partition. Is that s.o.p.? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e6f8d3db-0d5b-4b34-880e-25b977b4b...@halblog.com
Re: Command Line Partitioning
Honestly, that's WAY more of a level of involvement than what I want to get into. I've already figured out what I need to do and compared it to the size of the memory and the CPU and so on. I have a general rule: the less work (and still achieving my goal), the better. The fewer exceptions I need to maintain, the better. If all goes well, I'll be adding systems along the way, they're going to be in remote locations, and I don't want to be dealing with issues of custom kernels. I want the system to be as stock Debian as possible so I can get security updates and so on with apt-get update apt-get upgrade. While I know it's not always that easy, I've learned that the less I customize, the less there is that is likely to go wrong later. What I have found (as recommended by someone else) is sfdisk, which gives me the abilities of fdisk in a program intended for scripting (actually, reading the man page reminds me I could use input redirection for sfdisk and that might work for fdisk as well). It's a simple solution and doesn't create anything that'll add to the work of maintaining it later. Thank you for the suggestions, though. They are things I'll be looking into with other projects. Hal On Feb 19, 2011, at 1:02 PM, Stephen Powell wrote: On Sat, 19 Feb 2011 11:49:41 -0500 (EST), Hal Vaughan wrote: I have everything else worked out so it can be done in a batch mode or in a program, but I'm having trouble with partitioning. I've been working with cfdisk, fdisk, and parted: parted: Always leaves only 512 bytes at the start and grub-install doesn't seem to like that. But it can be used in a batch mode. fdisk: Doesn't seem to have a batch/script mode and I'd have to calculate sizes in megabytes from cylinder info cfdisk: Lets me work in whatever units I like (bytes, kbytes, mbytes, and so on) but is curses based and doesn't have batch/script mode. What I would like to do is to be able to use a partitioning program from a batch or script mode. Parted would be good for that, but I can't seem to get it to leave more than 512 bytes at the start (cfdisk left 32k -- why? I don't know). Also, it seems when I get CF disks they have an MS-DOS partition table (grub kept noting that and one part of the whole process didn't seem to like that). How can I either wipe the disk without a mess (tried zeroing it, parted wouldn't recognize it or write to it, but fdisk had no issue with that) or write a new partition table? What type of partition table should I use? (I'm guessing bsd, which was one of the options, but there wasn't a Linux option for types of partition tables.) Well, if I were in your situation I think I know what I would do. I would use another boot loader besides grub2, such as lilo or extlinux, that doesn't need to allocate sectors at the beginning of the disk which are supposed to be unallocated (or in the case of parted, are part of a partition). Both lilo and extlinux can be installed in either the master boot record or a partition boot sector and are perfectly happy with the MS-DOS partitioning scheme. With extlinux you are limited to either the ext2 or ext3 filesystem for the /boot partition. With lilo you can use just about anything for a file system. The down side of lilo is that the boot loader installer has to be run any time the physical location of any sector in the kernel image file or the initial RAM disk image file changes. I have a web page that you may find useful. It is http://users.wowway.com/~zlinuxman/Kernel.htm You may find it useful for several reasons. First of all, it is very common when working with embedded systems to create a custom Linux kernel. A lean and mean kernel is often created which contains only the drivers which are actually needed for that particular embedded system. This saves hard disk space and memory space on the embedded device (but not on the system on which one actually compiles the kernel!) Second, step 10, Customize the Kernel Installation Environment, contains stuff which is useful even to people who use stock kernels. For example, there is a section on how to switch to lilo from another boot loader (in this case, grub 1 from Lenny) and covers some little known stuff about lilo, such as how to make use of UUIDs in lilo, how to determine the capabilities of the BIOS, etc. Third, if you are using a custom kernel built in the manner recommended in the write-up, and you are using lilo, you will probably need some extra hook scripts which don't currently ship with lilo. This write-up provides some sample hook scripts which you may find useful. Good luck with your project. -- .''`. Stephen Powell : :' : `. `'` `- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org
Re: Cleaning DHCP and Host Info for New LAN
On Feb 17, 2011, at 6:40 AM, Bob wrote: On 02/17/2011 12:24 PM, Hal Vaughan wrote: On Feb 16, 2011, at 10:22 PM, Bob wrote: On 02/17/2011 06:38 AM, Hal Vaughan wrote: 8 snip system image pushed onto a CF card rm -f /mnt/src/etc/udev/rules.d/70-persistent-net.rules rm -f /mnt/src/etc/udev/rules.d/70-persistent-cd.rules Thanks for this one -- I don't have the cd.rules, but do have the net.rules and that would cause problems because of the different MAC addresses. Will the net.rules one be regenerated on boot if it doesn't exist? Yes. If you want a portable system where the LAN is always eth0 then you want to stop those files being generated, chmod -c 644 /lib/udev/write_* will do that but I get the impression with your project once the image is flashed onto a card it won't be moved that much. No, it won't be. In short, I'm working with a Soekris net5501, 512MB, 512Mhz with a slot for CF cards, a USB port, and an IDE port and PCI slot, but I'm not using IDE or PCI. I'm using the one I have now to develop the system and in the long run, I'll have several of these, each one going into a different office, which means different LANs. Once I get the image the way I want it on this CF card, I'll copy it and use it for other CF cards. When it's time to make a new unit, I'll copy the OS image onto a CF card, put that in the box, then put a USB RAM drive in. It'll find the file update.tgz, unpack it, and run a file within it. That file will use apt to first update the system, then add on more packages, and finally install the software I'm writing. So once I have a new CF card, it'll stay in one system and won't move at all, but the system will be put on a different LAN. rm -f /mnt/src/var/lib/ntp/ntp.drift Not even on this system. I will be installing ntp, but at a later stage. With my system images I try to configure as much as possible before I take the image as I have an NTP server on my firewall I like all my internal machines to sync off that which requires editing /etc/ntp.conf so I do it before and have the big block of commands I'm pasting onto a shell prompt sort it out for me. A lot of it's over kill, only some of my system images have swap file installed but it doesn't hurt to try delete a non-existent file this way the text file from which I paste the commands is more general. I'm using a minimal image on the CF card because I don't want to be changing the along the way, I want it stable and static. The install program in update.tgz will be where all the other configuration happens, since it'll be easier to make changes there and just re-tarball that image when I need to. Normally I'd agree with you and want the image to be as complete as possible. In this case, I think you could say I'm working with two images, the first is the one that goes onto the CF card and the other is the one including the update program and that's the one that will be complete when it's done. I can role out a nicely configured functional desktop image onto a harddrive, flash card, or USB stck in under 10 minutes which is very handy, after that all the system needs is updating (I have an apt-proxy so that's bloody quick) my friends old tired pox ridden winblows box is a sprightly Squeeze workstation. I can see how that would be quite useful, but other than these small embedded systems, I am not doing much in terms of installing or creating new systems. I use an iMac for my desktop now and my servers are on Linux. I also have a number of VMs under Parallels for work and on the last one, which was Squeeze, I just told Parallels where the iso was and it did the rest automatically. But VMs do have a lot of advantages over building something on new hardware. rm /etc/ssh/ssh_host_* dpkg-reconfigure openssh-server Almost all the ssh files will be re-configured, replaced, or deleted by the install program. I had not thought of reconfiguring. If the other files are deleted, what else does reconfiguring the ssh server do? (I'm actually considering not installing ssh until the update phase, since that'll force a new config for each system. But without ssh, if anything goes wrong, it's a pain to have to find the USB-serial adaptor and the cable and hook it all up to log in.) I do it to generate new keys it may not be the most elegant way but it works. Quick and simple, though. I like that because it eliminates work. After that it's mainly hardware specific stuff. Hardware isn't a real issue, other than MAC addresses, since I'm using a system where the board hasn't changed in a while and likely won't change for a good while, at least it likely won't change as long as I'm involved with this. So one system should look just like the rest hardware wise. That helps a lot, the 10 minutes quoted above doesn't take into account bullying ALSA to play nice and use the USB webcam mic by default etc.. Don't
Cleaning DHCP and Host Info for New LAN
I'm working with some embedded systems where the OS will be Squeeze on a CF card. The idea is to have a basic setup that will be stored in a zip file and when it's time to create a new system, the files will be unzipped onto a new CF card, then the card will be put in a new system. Of course to do this, I have to start with a base system or prototype and I don't want the info for my LAN or anything else like that to be passed on or left in the file system of the newer systems. When the new CF card is put into each new system and it is turned on, there's a command that will make the new system automatically download a config program that will make all needed changes to the embedded computer. That leads to the question, What do I change? I'll be changing the host name, which is in /etc/hostname and /etc/hosts. I also need to change the host name in /etc/exim4/update-exim4.conf.conf and in /etc/mailname. The user names in passwd and so on don't need changing, since each system has the same user name. I also know the domain is specified in /etc/resolv.conf. But if I'm getting the info for a network connection through DHCP, won't they be overwritten once the info comes in from the DHCP? Are there other files where I'd have to change the values so the new system doesn't contain references to the original LAN or host name or any other similar info? I'm not worried about files that will be re-written on the first boot (like anything dhclient or any DHCP stuff will re-write once it's on a new LAN), but I'd like to know if I'm missing any info that I need to erase or change. I'm not sure, but I think most of the info would be related to DNS or DHCP. Thanks for any info on this. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9e7efdbe-90af-467a-b24c-95c1ec816...@halblog.com
Re: Cleaning DHCP and Host Info for New LAN
On Feb 16, 2011, at 10:22 PM, Bob wrote: On 02/17/2011 06:38 AM, Hal Vaughan wrote: I'm working with some embedded systems where the OS will be Squeeze on a CF card. The idea is to have a basic setup that will be stored in a zip file and when it's time to create a new system, the files will be unzipped onto a new CF card, then the card will be put in a new system. Of course to do this, I have to start with a base system or prototype and I don't want the info for my LAN or anything else like that to be passed on or left in the file system of the newer systems. When the new CF card is put into each new system and it is turned on, there's a command that will make the new system automatically download a config program that will make all needed changes to the embedded computer. That leads to the question, What do I change? I'll be changing the host name, which is in /etc/hostname and /etc/hosts. I also need to change the host name in /etc/exim4/update-exim4.conf.conf and in /etc/mailname. The user names in passwd and so on don't need changing, since each system has the same user name. I also know the domain is specified in /etc/resolv.conf. But if I'm getting the info for a network connection through DHCP, won't they be overwritten once the info comes in from the DHCP? Are there other files where I'd have to change the values so the new system doesn't contain references to the original LAN or host name or any other similar info? I'm not worried about files that will be re-written on the first boot (like anything dhclient or any DHCP stuff will re-write once it's on a new LAN), but I'd like to know if I'm missing any info that I need to erase or change. I'm not sure, but I think most of the info would be related to DNS or DHCP. Thanks for any info on this. This does help! Some of these I'll be handling in specific ways. If there seems to be a problem with something I'm doing, feel free to let me know. when I copy my install I edit /etc/hostname That stays as is in the image so I can find it (and other files) with the host name and will be edited by my install program. rm -f /mnt/src/var/swap This doesn't exist, but I'm using a partition and not a swap file. It set that up by default and I don't think I need to change it. rm -f /mnt/src/etc/udev/rules.d/70-persistent-net.rules rm -f /mnt/src/etc/udev/rules.d/70-persistent-cd.rules Thanks for this one -- I don't have the cd.rules, but do have the net.rules and that would cause problems because of the different MAC addresses. Will the net.rules one be regenerated on boot if it doesn't exist? rm -f /mnt/src/var/lib/ntp/ntp.drift Not even on this system. I will be installing ntp, but at a later stage. rm /etc/ssh/ssh_host_* dpkg-reconfigure openssh-server Almost all the ssh files will be re-configured, replaced, or deleted by the install program. I had not thought of reconfiguring. If the other files are deleted, what else does reconfiguring the ssh server do? (I'm actually considering not installing ssh until the update phase, since that'll force a new config for each system. But without ssh, if anything goes wrong, it's a pain to have to find the USB-serial adaptor and the cable and hook it all up to log in.) After that it's mainly hardware specific stuff. Hardware isn't a real issue, other than MAC addresses, since I'm using a system where the board hasn't changed in a while and likely won't change for a good while, at least it likely won't change as long as I'm involved with this. So one system should look just like the rest hardware wise. Also see my response in the thread Installing Debian on USB sticks. Been looking at that, too -- thanks for the heads up on that! Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/04915edc-3445-4932-bfe3-98c38ac3c...@halblog.com
Trouble PXE Boot After Kernel Starts Loading
I have a Soekris Net5501 and I'm using tftp to boot pxelinux.0 and, ultimately, from there, I want to install Squeeze. First, I have to thank Debian, since the PXE boot explanation (sorry, lost the link from earlier) was a very good explanation on how to set up PXE. But since it was for Etch, I searched and also found an excellent one by Andy Smith here: http://andys.org.uk/wiki/Guide:PXE_network_booting_Debian_install that includes Squeeze. I think it should be pointed out that there is excellent documentation out there for this. Now, on to where I'm stuck, which is, I think, more a part of the normal boot process. I have the directory tree like this: tftpboot/debian/squeeze/i386/linux tftpboot/debian/squeeze/i386/initrd.gz tftpboot/pxelinux.cfg/default tftpboot/pxelinux.cfg/boot.txt It boots the kernel, but then I get: Probing EDD (edd=off to disable)... ok And after that, it freezes and I get nothing. The default config file, in full, is after my signature (it's longer, and I don't think it's the issue). I added edd=off to each kernel line in that config file because it was freezing up with the same message and suspected it was because (as best I know), there is no Enhanced Disk Drive. But then there's the ... ok after it, which makes me think the edd issue isn't why it's freezing. Is this due to the edd probe (in which case the edd=off isn't working for some reason)? If not, what should I check next? As best I know, while people have put all other versions of Debian on a Soekris Net5501, I haven't seen anything to tell me for sure people have yet successfully installed Squeeze on on. I'd rather not have to step back, since that would mean re-doing a number of systems later for updates. While that would not be for a while, of course, I'd like to be able to work from the latest version of Debian possible. Thanks for any insight into this. Hal --- tftpboot/pxelinux.cfg/default: --- DISPLAY boot.txt DEFAULT squeeze_i386_install LABEL squeeze_i386_install kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_linux kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_expert kernel debian/squeeze/i386/linux edd=off append priority=low vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_rescue kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz rescue/enable=true -- PROMPT 1 TIMEOUT 0 -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/227cca0f-d89d-412a-bf6d-920996163...@halblog.com
Re: Trouble PXE Boot After Kernel Starts Loading
On Feb 15, 2011, at 5:31 AM, Michael Tsang wrote: On Tuesday 15 February 2011 16:17:27 Hal Vaughan wrote: I have a Soekris Net5501 and I'm using tftp to boot pxelinux.0 and, ultimately, from there, I want to install Squeeze. First, I have to thank Debian, since the PXE boot explanation (sorry, lost the link from earlier) was a very good explanation on how to set up PXE. But since it was for Etch, I searched and also found an excellent one by Andy Smith here: http://andys.org.uk/wiki/Guide:PXE_network_booting_Debian_install that includes Squeeze. I think it should be pointed out that there is excellent documentation out there for this. Now, on to where I'm stuck, which is, I think, more a part of the normal boot process. I have the directory tree like this: tftpboot/debian/squeeze/i386/linux tftpboot/debian/squeeze/i386/initrd.gz tftpboot/pxelinux.cfg/default tftpboot/pxelinux.cfg/boot.txt It boots the kernel, but then I get: Probing EDD (edd=off to disable)... ok And after that, it freezes and I get nothing. The default config file, in full, is after my signature (it's longer, and I don't think it's the issue). I added edd=off to each kernel line in that config file because it was freezing up with the same message and suspected it was because (as best I know), there is no Enhanced Disk Drive. But then there's the ... ok after it, which makes me think the edd issue isn't why it's freezing. Is this due to the edd probe (in which case the edd=off isn't working for some reason)? If not, what should I check next? I think you should check the Linux documentation. EDD was OK but the next went wrong. Unfortunately, while I keep searching for the pxelinux boot sequence, I get instructions on how to set up pxe and the overall sequence, but nothing that breaks down what is actually done by the kernel. Is it transferring to the ram drive after this? I don't know -- there's not much on what the kernel does in a pxe boot sequence. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/d7b8bb71-86ea-4307-a8c7-00d121f83...@halblog.com
Re: Trouble PXE Boot After Kernel Starts Loading
This turned out to be something completely different and, it turns out, something VERY poorly documented. To communicate with the Soekris system, I need to use a serial cable and the program screen. It turns out that the baud rate is changed during the install unless you specify a baud rate in 3 different places. What was happening was that when the baud rate changed, rather than getting garbage on the screen, I was getting nothing, so it didn't seem like a baud issue. Hal On Feb 15, 2011, at 3:17 AM, Hal Vaughan wrote: I have a Soekris Net5501 and I'm using tftp to boot pxelinux.0 and, ultimately, from there, I want to install Squeeze. First, I have to thank Debian, since the PXE boot explanation (sorry, lost the link from earlier) was a very good explanation on how to set up PXE. But since it was for Etch, I searched and also found an excellent one by Andy Smith here: http://andys.org.uk/wiki/Guide:PXE_network_booting_Debian_install that includes Squeeze. I think it should be pointed out that there is excellent documentation out there for this. Now, on to where I'm stuck, which is, I think, more a part of the normal boot process. I have the directory tree like this: tftpboot/debian/squeeze/i386/linux tftpboot/debian/squeeze/i386/initrd.gz tftpboot/pxelinux.cfg/default tftpboot/pxelinux.cfg/boot.txt It boots the kernel, but then I get: Probing EDD (edd=off to disable)... ok And after that, it freezes and I get nothing. The default config file, in full, is after my signature (it's longer, and I don't think it's the issue). I added edd=off to each kernel line in that config file because it was freezing up with the same message and suspected it was because (as best I know), there is no Enhanced Disk Drive. But then there's the ... ok after it, which makes me think the edd issue isn't why it's freezing. Is this due to the edd probe (in which case the edd=off isn't working for some reason)? If not, what should I check next? As best I know, while people have put all other versions of Debian on a Soekris Net5501, I haven't seen anything to tell me for sure people have yet successfully installed Squeeze on on. I'd rather not have to step back, since that would mean re-doing a number of systems later for updates. While that would not be for a while, of course, I'd like to be able to work from the latest version of Debian possible. Thanks for any insight into this. Hal --- tftpboot/pxelinux.cfg/default: --- DISPLAY boot.txt DEFAULT squeeze_i386_install LABEL squeeze_i386_install kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_linux kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_expert kernel debian/squeeze/i386/linux edd=off append priority=low vga=normal initrd=debian/squeeze/i386/initrd.gz -- LABEL squeeze_i386_rescue kernel debian/squeeze/i386/linux edd=off append vga=normal initrd=debian/squeeze/i386/initrd.gz rescue/enable=true -- PROMPT 1 TIMEOUT 0 -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/227cca0f-d89d-412a-bf6d-920996163...@halblog.com -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/f019d40b-00ce-4c7f-a2b2-9573d4f21...@halblog.com
Re: Why is Evolution and Epiphany now a part of gnome-core?
On Jan 17, 2011, at 12:16 PM, T o n g wrote: On Sun, 02 Jan 2011 13:08:36 -0800, Mike Bird wrote: I'm running Debian Squeeze, and I have only gnome-core installed so that I would not have Evolution or Epiphany installed since I do not use them. Me too. Basically because gnome-desktop-environment is too big to fit on CD 1. See bug #608098 for more information, especially http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=608098#31. That seems to me an absurd reason. So only 3 of us in the entire Debian world think this way? Re-quoting OP of bug 608098: ,- | The massive migration of dependencies from | gnome-desktop-environment to gnome-core is extremely undesirable, | because it spoils the usefulness that gnome-core used to have in | pulling just enough packages to have a basic GNOME | environment. Now, instead, it pulls WAY too many packages and | leaves the user without any simple method for installing basic | GNOME components. `- I still think it make perfect sense, and a legitimated request. Creating a new package to depend upon evolution and ephiphany and gnome-core would be a less harmful solution. Yeah, but look at what the maintainer said: ,- | The gnome-core package is not here to fulfill the needs of a given | user. | | If you need a specific set of packages, please make your metapackages | yourself. `- well... what I can do, huh? Nobody care about this? This is an example of why I've been moving away from FOSS. Someone makes a good point in a bug report and the programmer/developer/maintainer throws it back in his face, which allows the bug to be closed out quickly. I think it's a legitimate concern and just because one person pointed it out does not mean it only effects one person. Yes, I care about it. I usually run Debian for headless systems that don't use X or a DE, but when I'm using a GUI on Debian, and need to use a Gnome program, there's much more sense to it requiring and installing the bare minimum of what it needs than installing a lot of bloat. Isn't the Debian way more about allowing customization and forcing as little as possible on the users or sys-admins? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/d0d337f4-b065-4f51-94e3-2d0c2e16e...@halblog.com
Re: Why is Evolution and Epiphany now a part of gnome-core?
On Jan 17, 2011, at 12:48 PM, Carl Fink wrote: On Mon, Jan 17, 2011 at 12:26:17PM -0500, Hal Vaughan wrote: This is an example of why I've been moving away from FOSS. Someone makes a good point in a bug report and the programmer/developer/maintainer throws it back in his face, which allows the bug to be closed out quickly. I think it's a legitimate concern and just because one person pointed it out does not mean it only effects one person. It's ridiculous, and as you write unfortunately typical. I recall getting similar Sorry, I don't care stuff from developers as far back as when StarOffice was first freed. What to do? Well, in my case, I switch to different FOSS. So ... I guess I move to something other than Debian. Oh well. I rarely file bug reports for FOSS anymore due to responses like this. But I have started communicating with people in some projects when something is wrong (either on the forum or on mailing lists). I tried to explain an issue when I filed a bug with OOo about margins, but the responses showed they were not concerned with listening to the issue, only with saying, We're right, so go away. At one point I was trying to design a web interface that would be easy to customize for LIRC so you could control your home devices through LIRC from a webpage you could pull up on your smart phone or on a PDA and asked for help. I got nothing. I asked again, knowing that the people who were behind LIRC were on the list. I finally left a message saying, I wanted to add to the project and asked for help several times and was ignored or told to RTFM when I had stated the info wasn't in TFM. I've scrapped my project and LIRC devs might want to consider not ignoring help requests in the future. I watched and saw everyone say, Well, he must have asked for help wrong. (Hey, I've read ESR's How to Ask for Help and found most of it common sense and I follow it.) Not one looked for my original requests to see if I was rude or anything. I've been in discussions on MacPorts for KDE on OS X and found some helpful people, but I have to say some devs are too busy being right to listen. I could list more, but I've seen enough cases where bug reports are met with dismissals (and often rude or condescending ones, like the one cited early in this thread) that I don't file bug reports for FOSS unless I know the project and know it's worth the time to file one. I think this is a classic case of some devs being so into themselves and their projects that they don't realize they come across as Sheldon Cooper when they ignore others. I retired at 45, thanks to my business, which was based on my own custom software. The first program I released to run on my clients' computers was in Java, within 6 months after I learned Java and OOP. In the next 18 months that version was in use, I had fewer than 5 bug reports (because I tested the hell out of it before releasing it). But I had to respond to each bug report and usually had the fix out within 12 hours. I would have lost a lot of money if I had been as dismissive of bug reports as some developers are. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/0d534540-eabb-498a-b859-30386e01b...@halblog.com
DD To a Smaller Partition
I have a 10 GB partition that is nowhere near full, less than 5 GB of data on it (far less). Unfortunately, the partition I'd like to copy it to is 5 GB. I can do rsync -av but normally I'd use DD. Is there a way to copy an image of just the files from one drive to another? For example, an intelligent DD that won't copy unused sectors, or will create an archive of just the files? Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/abf06b01-fe96-44da-b00d-fb16ee6e8...@halblog.com
Where Is the List of Installed Packages?
I have a frustrating situation. I had a RAID go out on me and require rebuilding. It's mostly rebuilt and all my data is intact, but not all the settings and config info and program files have been copied over yet. And, while I have to rebuild that RAID, which is my back up -- well, you can predict the rest. That's when a system went down on me. It was sometime last night and now I can't even get video on the monitor for a POST, so the mobo is fried. I do have a system I can transfer this one to, however, there's one problem: The new system uses SATA and the old drivers are PATA. Fortunately, I can run a PATA drive on the new system, but it's going to be easier if I don't boot off it for now. I can simply re-install Debian with no trouble, the only issue is that I don't have the up-to-date list of all the packages installed on the old PATA drive. I can, though, plug it in and read it as a data drive. So where on that drive can I find the list of installed packages? Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2e72e6f5-66cb-45cb-a2f8-0f6ec706a...@halblog.com
Re: Where Is the List of Installed Packages?
On Jan 8, 2011, at 12:58 PM, Andrei Popescu wrote: On Sb, 08 ian 11, 18:42:53, David Sastre wrote: On Sat, 08 Jan 2011 10:22:01 -0500, Hal Vaughan wrote: I can simply re-install Debian with no trouble, the only issue is that I don't have the up-to-date list of all the packages installed on the old PATA drive. I can, though, plug it in and read it as a data drive. So where on that drive can I find the list of installed packages? You could also mount that PATA drive externally and chroot into it to request that (or any other) info: # dpkg -l list.of.packages.txt The result is a much smaller file: # dpkg -l list If you can/want to use dpkg better use --get-selections because the output can be fed to --set-selections ;) Without doing a chroot (while I know the mobo on the old system crashed, I'm beginning to suspect drive/OS issues), I could use a few utils like grep and awk to change a list from /var/lib/dpkg/status if I needed to, couldn't I? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/58ac1d4b-f02c-475d-8fff-025594a20...@halblog.com
Re: Where Is the List of Installed Packages?
Thank you, everyone, for the answers. This will make restoring this go faster. Just a little note with this to say I can't tell people on this list and the Debian Developers how much I appreciate Debian. Right now I'm having issues with the new hardware fitting in place and irritating stuff like that, which is dragging things out. But since this is on Debian, experience tells me that once the hardware is all working, setting up the replacement system (even if I have to do it from scratch) will not take long because of all the careful planning and work that goes into building Debian. I'll be able to use the same packages I had before, altering the configurations is almost unneeded in most cases (I think all I have to change is sshd_config for RSA auth. and no password auth and change MySQL so it uses the directories on my RAID instead of the defaults). So thank you for the help on this, but overall, thank you for Debian and all those willing to help on this list. Hal On Jan 8, 2011, at 10:22 AM, Hal Vaughan wrote: I have a frustrating situation. I had a RAID go out on me and require rebuilding. It's mostly rebuilt and all my data is intact, but not all the settings and config info and program files have been copied over yet. And, while I have to rebuild that RAID, which is my back up -- well, you can predict the rest. That's when a system went down on me. It was sometime last night and now I can't even get video on the monitor for a POST, so the mobo is fried. I do have a system I can transfer this one to, however, there's one problem: The new system uses SATA and the old drivers are PATA. Fortunately, I can run a PATA drive on the new system, but it's going to be easier if I don't boot off it for now. I can simply re-install Debian with no trouble, the only issue is that I don't have the up-to-date list of all the packages installed on the old PATA drive. I can, though, plug it in and read it as a data drive. So where on that drive can I find the list of installed packages? Thank you! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2e72e6f5-66cb-45cb-a2f8-0f6ec706a...@halblog.com -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/a36dea80-5ae5-477a-b59b-9ab48e667...@halblog.com
Re: [SOLVED!] Communicating with USB Modem
I was going to just reply to Phil privately, but there are a few points that could effect someone else's decision in a similar case, so I'm going to address them on list. Phil makes some good points, so I thought it appropriate to include my thoughts on them. On Oct 18, 2010, at 12:38 AM, Phil Requirements wrote: On 2010-10-13 22:28:10 -0400, Hal Vaughan wrote: Comments and some info at the bottom, so it makes sense when you read it. (Hey, there's NO way I'm going to top post on this list!) Now I have a few notes on this, since there's been a few private conversations with people who are in or have seen this thread. 1) I'm basically making a black box for the people I'm working with. I do NOT want them messing with the computer I give them. Many times, in embedded systems, there is no PCI slot. Sometimes there is, but I can't count on that. (I haven't picked my hardware config yet.) So I really need an external modem since the only connection I can count on is a USB connector. (Even with embedded, most systems have 2 or more USB connectors, and some people may need to hook up a printer there.) I see that here in Point 1 you rule out a PCI card-based solution while also admitting that the hardware has not yet been specified. It seems feasible that if you found a really great and cheap PCI modem, you could spec the hardware around that. I spent close to a week researching possible mobos or all-in-one package systems. Some boards do have PCI slots, but some don't, so that's a big issue to consider. I'd much rather be able to spec the hardware around the mobo than a modem. Also, if I go for something like Soekris, which I like, but which may cost more than I want (they still haven't released final specs and cost on their Net6501, which is what I'd likely use), using a PCI card would be problematical. While I've seen PCI slots on their boards, I'd have to cut a hole in it to allow someone to plug a phone cable in. That's one of the two biggest strikes against a PCI modem: I can't be sure how it'll fit in with the rest, and finding an appropriate mobo or mobo/case combo is a bigger concern than using a PCI modem. If I go in order of my reasoning, I listed that as strike 2 against PCI modems. I mention this because it seems to me that an external USB modem is not very black box. There would be your discrete little appliance with a very conspicuous dongle coming off it. A PCI internal modem that fits inside the appliance would be much more black-box-like. I agree. Until I read how you phrased it here, I had not thought of stating my reasoning here. I'll be working intelligent people who work in technical fields, but not in computer fields. The tradeoff is that if I need someone in another city to work with me so I can walk them through solving a problem so I don't have to make 4-10 hour round trip to get there and end up just flipping a switch, that's good, since they can work with me. But, on the flip side, when you're working with someone with an intelligent and technical mind, they're the ones more likely to think they can break the rules in some way because they know they're smart. I'm more concerned about this being black boxy in terms of them not trying to log on or hook up a keyboard and monitor than anything else. I don't want them thinking of this box as a computer where, if it starts acting odd, they think they can work with. On the other hand, even though I'm within 8 miles of the state capital building, I didn't get broadband for years, due to bureaucratic attitudes at the cable company (which kept changing ownership) and Verizon. I had to work with Linux and dial-up modems for several years. I also had to replace my PCI modems every few months. I don't remember why, but I do remember tossing out a lot of PCI modems that worked and stopped working. Sometimes it was due to a distro upgrade, other times it was a new modem that docs said would work that didn't. If a PCI modem goes bad on me, then it means paying for them to ship me the unit overnight, me replacing it and sending it back overnight. If a USB modem goes bad, it means having a new one drop-shipped from Newegg. Considering package size, that's much easier and cheaper than replacing a PCI modem. That, on my list, was strike 3 against PCI. The problem with PCI card modems is that they are not so likely to support GNU/Linux. So they aren't a clear winner. But they would be better for a black box. And that was my strike 1 against PCI -- along with seeing some work now and not with later drivers. As I said, my black box was a bit of a compromise. It's a black box as far as being something a user can log onto, but not 100% in that it has the external modem. Still, what I'm looking at now includes the embedded box, a PCI modem, and a wall-wart transformer. All the user has to do on receiving it is plug in a CAT5, plug in the USB modem, plug
Re: [SOLVED!] Communicating with USB Modem
Comments and some info at the bottom, so it makes sense when you read it. (Hey, there's NO way I'm going to top post on this list!) On Oct 9, 2010, at 3:21 AM, Hal Vaughan wrote: I have a Debian Lenny system and I've plugged in a USB 56K modem. (I know that's as outdated as a Model T, but I need it for business.) When I type cat /proc/bus/usb/devices I get this: T: Bus=02 Lev=00 Prnt=00 Port=00 Cnt=00 Dev#= 1 Spd=480 MxCh= 8 B: Alloc= 0/800 us ( 0%), #Int= 0, #Iso= 0 D: Ver= 2.00 Cls=09(hub ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1d6b ProdID=0002 Rev= 2.06 S: Manufacturer=Linux 2.6.26-2-686 ehci_hcd S: Product=EHCI Host Controller S: SerialNumber=:00:1d.7 C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr= 0mA I:* If#= 0 Alt= 0 #EPs= 1 Cls=09(hub ) Sub=00 Prot=00 Driver=hub E: Ad=81(I) Atr=03(Int.) MxPS= 4 Ivl=256ms T: Bus=01 Lev=00 Prnt=00 Port=00 Cnt=00 Dev#= 1 Spd=12 MxCh= 2 B: Alloc= 0/900 us ( 0%), #Int= 0, #Iso= 0 D: Ver= 1.10 Cls=09(hub ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1d6b ProdID=0001 Rev= 2.06 S: Manufacturer=Linux 2.6.26-2-686 uhci_hcd S: Product=UHCI Host Controller S: SerialNumber=:00:1d.0 C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr= 0mA I:* If#= 0 Alt= 0 #EPs= 1 Cls=09(hub ) Sub=00 Prot=00 Driver=hub E: Ad=81(I) Atr=03(Int.) MxPS= 2 Ivl=255ms T: Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 2 Spd=12 MxCh= 0 D: Ver= 2.00 Cls=ff(vend.) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=047e ProdID=2892 Rev= 2.00 S: Manufacturer=Agere Systems S: Product=Agere USB2.0 V.92 SoftModem C:* #Ifs= 1 Cfg#= 1 Atr=80 MxPwr=200mA I:* If#= 0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=82(I) Atr=01(Isoc) MxPS= 0 Ivl=1ms E: Ad=02(O) Atr=01(Isoc) MxPS= 0 Ivl=1ms I: If#= 0 Alt= 1 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=82(I) Atr=01(Isoc) MxPS= 32 Ivl=1ms E: Ad=02(O) Atr=01(Isoc) MxPS= 32 Ivl=1ms I: If#= 0 Alt= 2 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=81(I) Atr=02(Bulk) MxPS= 64 Ivl=0ms E: Ad=01(O) Atr=02(Bulk) MxPS= 64 Ivl=0ms E: Ad=82(I) Atr=03(Int.) MxPS= 16 Ivl=16ms From what I can find, that means the actual device file for this modem should be at /dev/bus/usb/001/002, but I know I could be wrong on that. I'm trying to communicate with this modem. I've used Minicom, with the device I mentioned above. There are no /dev/ttyACM0 devices, as sometimes happens with a modem. I've tried all /dev/ttySx, where x is from 0 to 4, with no results. The problem with Minicom is that it can take time and a lot of frustrating keypresses to try one device after another. Also, I'm not sure the serial port speed settings are at all appropriate in Minicom for USB modems. I was also trying a Perl program that would simply fork, have the parent open the device for reading and listen, then the child would open the same device for writing and send ATZ through to the device, then I could watch for a response -- but then when I tried it on a system with a serial port where I knew the device was /dev/ttyS0, it didn't work, so maybe there's a problem with the program. (I don't want to post a full Perl program here, since this isn't a Perl board that focuses on debugging Perl.) This is a test situation, so I can't be sure the modem works with Linux, but I would think if it's external and USB, it would have to accept serial commands sent over the USB port. I could be wrong on this, of course. So here's the questions: 1) What device should I be able to write to and read from to connect to this modem? 2) Any suggestions on how to write to and read from the device? (Write to it in Perl or use echo in bash -- but then how do I read to see results?) 3) Any other ideas or suggestions on how to control and use this modem in Linux (without going into C or C++)? Even after playing around for a while, and with help from others (see other emails in this thread), I could not get the other modem to work properly. I did try the sl-modem drivers from the non-free repositories. They would create /dev/modem and link it to /dev/ttySL0, but even with that, it wouldn't work. When I tested any modem, I did two things: 1) I used Minicom, which is a great term program, and 2) I used my own Perl program that would open a device as a file to read and write to, then send the ATZ command (with a newline character after it) and I'd see if anything came back. I missed a few details on this program, but once I got it to work (tested it on an RS232 modem on another box), I would see the ATZ printed on the console (as well as knowing it was supposedly sent to the device) and I'd see a blank line, then OK. No matter what I did with Minicom or my Perl script, I could not communicate with the Encore modem. Today I received another USB modem, a Rosewill RNX-56USB from NewEgg. The link to that item is: http://www.newegg.com/Product
Communicating with USB Modem
I have a Debian Lenny system and I've plugged in a USB 56K modem. (I know that's as outdated as a Model T, but I need it for business.) When I type cat /proc/bus/usb/devices I get this: T: Bus=02 Lev=00 Prnt=00 Port=00 Cnt=00 Dev#= 1 Spd=480 MxCh= 8 B: Alloc= 0/800 us ( 0%), #Int= 0, #Iso= 0 D: Ver= 2.00 Cls=09(hub ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1d6b ProdID=0002 Rev= 2.06 S: Manufacturer=Linux 2.6.26-2-686 ehci_hcd S: Product=EHCI Host Controller S: SerialNumber=:00:1d.7 C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr= 0mA I:* If#= 0 Alt= 0 #EPs= 1 Cls=09(hub ) Sub=00 Prot=00 Driver=hub E: Ad=81(I) Atr=03(Int.) MxPS= 4 Ivl=256ms T: Bus=01 Lev=00 Prnt=00 Port=00 Cnt=00 Dev#= 1 Spd=12 MxCh= 2 B: Alloc= 0/900 us ( 0%), #Int= 0, #Iso= 0 D: Ver= 1.10 Cls=09(hub ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1d6b ProdID=0001 Rev= 2.06 S: Manufacturer=Linux 2.6.26-2-686 uhci_hcd S: Product=UHCI Host Controller S: SerialNumber=:00:1d.0 C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr= 0mA I:* If#= 0 Alt= 0 #EPs= 1 Cls=09(hub ) Sub=00 Prot=00 Driver=hub E: Ad=81(I) Atr=03(Int.) MxPS= 2 Ivl=255ms T: Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 2 Spd=12 MxCh= 0 D: Ver= 2.00 Cls=ff(vend.) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=047e ProdID=2892 Rev= 2.00 S: Manufacturer=Agere Systems S: Product=Agere USB2.0 V.92 SoftModem C:* #Ifs= 1 Cfg#= 1 Atr=80 MxPwr=200mA I:* If#= 0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=82(I) Atr=01(Isoc) MxPS= 0 Ivl=1ms E: Ad=02(O) Atr=01(Isoc) MxPS= 0 Ivl=1ms I: If#= 0 Alt= 1 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=82(I) Atr=01(Isoc) MxPS= 32 Ivl=1ms E: Ad=02(O) Atr=01(Isoc) MxPS= 32 Ivl=1ms I: If#= 0 Alt= 2 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none) E: Ad=81(I) Atr=02(Bulk) MxPS= 64 Ivl=0ms E: Ad=01(O) Atr=02(Bulk) MxPS= 64 Ivl=0ms E: Ad=82(I) Atr=03(Int.) MxPS= 16 Ivl=16ms From what I can find, that means the actual device file for this modem should be at /dev/bus/usb/001/002, but I know I could be wrong on that. I'm trying to communicate with this modem. I've used Minicom, with the device I mentioned above. There are no /dev/ttyACM0 devices, as sometimes happens with a modem. I've tried all /dev/ttySx, where x is from 0 to 4, with no results. The problem with Minicom is that it can take time and a lot of frustrating keypresses to try one device after another. Also, I'm not sure the serial port speed settings are at all appropriate in Minicom for USB modems. I was also trying a Perl program that would simply fork, have the parent open the device for reading and listen, then the child would open the same device for writing and send ATZ through to the device, then I could watch for a response -- but then when I tried it on a system with a serial port where I knew the device was /dev/ttyS0, it didn't work, so maybe there's a problem with the program. (I don't want to post a full Perl program here, since this isn't a Perl board that focuses on debugging Perl.) This is a test situation, so I can't be sure the modem works with Linux, but I would think if it's external and USB, it would have to accept serial commands sent over the USB port. I could be wrong on this, of course. So here's the questions: 1) What device should I be able to write to and read from to connect to this modem? 2) Any suggestions on how to write to and read from the device? (Write to it in Perl or use echo in bash -- but then how do I read to see results?) 3) Any other ideas or suggestions on how to control and use this modem in Linux (without going into C or C++)? Thanks for any help on this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8f2308b6-5559-40d6-a963-dd1582010...@halblog.com
Re: Communicating with USB Modem
On Oct 9, 2010, at 7:24 AM, Camaleón wrote: On Sat, 09 Oct 2010 03:21:36 -0400, Hal Vaughan wrote: I have a Debian Lenny system and I've plugged in a USB 56K modem. (I know that's as outdated as a Model T, but I need it for business.) When I type cat /proc/bus/usb/devices I get this: (...) As Ron suggested, better look at dmesg messages. - If the device is not detected, it will (not) be there. - If the device is detected but has problems for being setup, it will be there. - If the device is properly detected and configured, it will also be there. From what I can find, that means the actual device file for this modem should be at /dev/bus/usb/001/002, but I know I could be wrong on that. USB modems should fall under /dev/ttyUSB0 There's no /dev/ttyUSB0. Just the usual /dev/tty and the same with 0-63 on the end and /dev/ttyS0 through S3. I tried all of them and all the possible USB devices in that directory. I'm trying to communicate with this modem. I've used Minicom, with the device I mentioned above. There are no /dev/ttyACM0 devices, as sometimes happens with a modem. I've tried all /dev/ttySx, where x is from 0 to 4, with no results. The problem with Minicom is that it can take time and a lot of frustrating keypresses to try one device after another. Also, I'm not sure the serial port speed settings are at all appropriate in Minicom for USB modems. Don't go nuts with this. Just plug the modem, open a console, type dmesg | grep -i usb and put the result here (for very long logs, upload the result to www.pastebin.com or any service like that) :-) As you and Ron suggested. It's the last device. No errors, and remember it showed up in /proc/bus/usb/devices and looked okay (I numbered the lines in the output from grep): 235:[1.849802] usbcore: registered new interface driver usbfs 236:[1.849824] usbcore: registered new interface driver hub 237:[1.849848] usbcore: registered new device driver usb 238:[1.853655] USB Universal Host Controller Interface driver v3.0 243:[1.854354] uhci_hcd :00:1d.0: new USB bus registered, assigned bus number 1 245:[1.854438] usb usb1: configuration #1 chosen from 1 choice 246:[1.854455] hub 1-0:1.0: USB hub found 251:[1.956563] usb usb1: New USB device found, idVendor=1d6b, idProduct=0001 252:[1.956566] usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1 253:[1.956567] usb usb1: Product: UHCI Host Controller 254:[1.956569] usb usb1: Manufacturer: Linux 2.6.26-2-686 uhci_hcd 255:[1.956570] usb usb1: SerialNumber: :00:1d.0 260:[1.956976] ehci_hcd :00:1d.7: new USB bus registered, assigned bus number 2 263:[1.968481] ehci_hcd :00:1d.7: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004 264:[1.968538] usb usb2: configuration #1 chosen from 1 choice 265:[1.968555] hub 2-0:1.0: USB hub found 269:[2.072362] usb usb2: New USB device found, idVendor=1d6b, idProduct=0002 270:[2.072364] usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1 271:[2.072366] usb usb2: Product: EHCI Host Controller 272:[2.072367] usb usb2: Manufacturer: Linux 2.6.26-2-686 ehci_hcd 273:[2.072369] usb usb2: SerialNumber: :00:1d.7 281:[2.196626] usb 1-1: new full speed USB device using uhci_hcd and address 2 282:[2.347904] usb 1-1: not running at top speed; connect to a high speed hub 283:[2.352211] usb 1-1: configuration #1 chosen from 1 choice 285:[2.485485] usb 1-1: New USB device found, idVendor=047e, idProduct=2892 286:[2.485488] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0 287:[2.485490] usb 1-1: Product: Agere USB2.0 V.92 SoftModem 288:[2.485491] usb 1-1: Manufacturer: Agere Systems Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/39759fa6-1dfb-46ea-93cc-a2572dddc...@halblog.com
Re: Communicating with USB Modem
On Oct 9, 2010, at 12:44 PM, Camaleón wrote: On Sat, 09 Oct 2010 11:39:49 -0400, Hal Vaughan wrote: On Oct 9, 2010, at 7:24 AM, Camaleón wrote: Don't go nuts with this. Just plug the modem, open a console, type dmesg | grep -i usb and put the result here (for very long logs, upload the result to www.pastebin.com or any service like that) :-) As you and Ron suggested. It's the last device. No errors, and remember it showed up in /proc/bus/usb/devices and looked okay (I numbered the lines in the output from grep): (...) 285:[2.485485] usb 1-1: New USB device found,idVendor=047e, idProduct=2892 286:[2.485488] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0 287:[2.485490] usb 1-1: Product: Agere USB2.0 V.92 SoftModem 288:[2.485491] usb 1-1: Manufacturer: Agere Systems There you got it. It's an Agere soft modem. Bufff... as per this doc¹ you could try sl-modem package from non-free repo (it seems that your device -SV92U2- uses the Scorpio chipset) but prepare for the worst ;-( OTOH, LSI (the owner company of Agere) states that the modem supports Linux, so you can ask them for a driver :-? ¹ http://www.modemsite.com/56k/lucentamr.asp I had a reference, and now I can't find the darned link, to it working on Linux, otherwise I wouldn't have bought it -- unless I had a bunch of tabs open when I was researching modems at Newegg and hit add to cart on the wrong one -- which could happen. I can't remember where (so it could have been the manufacturer's site), but there was one place I read that if it's USB, it had to use the Hayes command set and would work on anything, but Brad's link to TLDP shows that wrong. Next time when I see positives on something like that, I'll look for the flip side, just in case. I've already ordered one from NewEgg that has several reviewers saying they're using it on Linux. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/bb12d9f1-60d7-4f9b-8758-47043ea7a...@halblog.com
Re: Communicating with USB Modem
On Oct 9, 2010, at 4:22 PM, Camaleón wrote: On Sat, 09 Oct 2010 15:29:40 -0400, Hal Vaughan wrote: On Oct 9, 2010, at 12:44 PM, Camaleón wrote: (...) Bufff... as per this doc¹ you could try sl-modem package from non-free repo (it seems that your device -SV92U2- uses the Scorpio chipset) but prepare for the worst ;-( OTOH, LSI (the owner company of Agere) states that the modem supports Linux, so you can ask them for a driver :-? ¹ http://www.modemsite.com/56k/lucentamr.asp I had a reference, and now I can't find the darned link, to it working on Linux, otherwise I wouldn't have bought it -- unless I had a bunch of tabs open when I was researching modems at Newegg and hit add to cart on the wrong one -- which could happen. I can give you at least one reference (the manufacture's tech. specs): *** http://www.agere.com/docs/PCS_Catalog_052606.pdf (page 5) – SV92U2: USB 2.0 device controller, 48-pin TQFP Stack Bus (...) OS support: Win98SE/2000/ME, WinXP, WinXP 64-bit, Vista (Native Support), and Linux *** I see that, the one thing that is discouraging me is that, while I know they are talking about the chip, the pictures are of PCI modems, not the USB ones, and I think mine is later. (They show a USB modem on page 5, but different form factor than mine.) Still, mine may work. If that is indeed your device, don't give up so easily :-) Just try with sl-modem drivers or ask LSI/Agere for advice, it could work. I tried the page you linked to, which had a link to the site with drivers, but that 2nd site is all blank web pages -- still looking for the sl-modem drivers. Did I miss a link? That's possible. (The dead link is: http://www.smlink.com/. It's from the last paragraph of that page, the only section that covers Linux there.) I can't remember where (so it could have been the manufacturer's site), but there was one place I read that if it's USB, it had to use the Hayes command set and would work on anything, but Brad's link to TLDP shows that wrong. Next time when I see positives on something like that, I'll look for the flip side, just in case. I've already ordered one from NewEgg that has several reviewers saying they're using it on Linux. When it comes to modems and linux, the only way to hit the right device is by using a RS-232 modem. No drivers needed and straight-forward setup for all kind of services (dial-up connection, fax facility...). Yes, yes... I know. Serial port is a scarce resource in modern motherboards and n[eo]tbook computers ;-( A little backstory here. My small business mines data, and a lot comes from some dial-up systems where you can get passwords, but most people ignore them now since they're harder to deal with (therefore, if you're bigger, less lucrative, if you're smaller, good money). I have a US Robotics RS-232 running on the main system now and I have another that was brand new, that I pulled out and tested, then put back in the box. I guard those modems carefully. They work and I know they work. But an idea hit me this summer. I haven't been programming in about 3 years (I'm a screenwriter by passion, so I've been doing that). This idea that hit me would take several months of programming and it'd be a game-changer for me in terms of income from the data mining. But to do it, I have to decentralize, which means instead of having one computer here doing dial-up, I'd have to put a computer in each client's office and have it do dial-up there. I won't get into all the reasons and thinking behind what I'm doing, but, in short, I want what's in their offices to be as simple as possible and to be a black box. I do not want them hooking up a keyboard or monitor to it ever. I don't even want them to think of it as a computer, EVER! So I'm looking not at low end as in cheap, but as in saving money and still getting a good embedded system. Each system will need a modem and I'll keep backups (of the computer and modem) on hand, ready to ship when needed. I have used a USB-to-RS-232 converter with success for a FOSS project to control an HD radio (http://halblog.com/hdradiocontroller.html), but an RS-232 modem is more expensive and the converters are expensive, too, putting the price per modem well over $50 each. I've ordered a Rosewill that looks good. It costs more than the one I have now. I'm going to try any drivers I can find for this one, but considering the price of the other one, with shipping, is $30, if the drivers don't work and the manufacturer doesn't help, I'm not going to knock myself out over this one. I can always use it on a Windows machine and use that system for testing (to have my other ones dial in). As it is, though, considering how cheap these are (this one and the replacement people have tried on Ubuntu), I think spending more than a few hours on this one might be a diminishing return. Still, if all it needs are drivers
Re: Communicating with USB Modem
On Oct 9, 2010, at 6:49 PM, Camaleón wrote: On Sat, 09 Oct 2010 17:28:31 -0400, Hal Vaughan wrote: On Oct 9, 2010, at 4:22 PM, Camaleón wrote: ... I see that, the one thing that is discouraging me is that, while I know they are talking about the chip, the pictures are of PCI modems, not the USB ones, and I think mine is later. (They show a USB modem on page 5, but different form factor than mine.) Still, mine may work. Yep, if you carefully read that doc, it seems the shape of the device may vary: *** - Reference design will include board, connectors, USB cable, and plastics - Three different form factors *** Yeah, I missed it. When I start researching, I look at everything, then as I focus on some parts, I am not as good about making sure I read all of it. Oops! ... Oops, sorry. I thought I already told you sl-modem drivers are available under debian non-free repo (lenny, squeeze, sid): http://packages.debian.org/lenny/sl-modem-daemon :-) I found that just after writing my last email on this thread, but just before I had to leave to take my Mother to night out meal - didn't have time to add in that I had found it. Thanks. The issue, though, is that needs another package, one that's not a dependency, it's sl-modem-modules-2.6.26-2-686. I'm looking, can even find it in a package pole, but can't find the package itself. ... But an idea hit me this summer. I haven't been programming in about 3 years (I'm a screenwriter by passion, so I've been doing that). This idea that hit me would take several months of programming and it'd be a game-changer for me in terms of income from the data mining. But to do it, I have to decentralize, which means instead of having one computer here doing dial-up, I'd have to put a computer in each client's office and have it do dial-up there. I won't get into all the reasons and thinking behind what I'm doing, but, in short, I want what's in their offices to be as simple as possible and to be a black box. I do not want them hooking up a keyboard or monitor to it ever. I don't even want them to think of it as a computer, EVER! So I'm looking not at low end as in cheap, but as in saving money and still getting a good embedded system. Each system will need a modem and I'll keep backups (of the computer and modem) on hand, ready to ship when needed. I see... It's quite difficult to keep the track of manufacturers making cheap and linux friendly devices. One has to perform a big search in the web (reading forums, mailing lists, asking users to get accurate feedback...) to find out what USB modems (chipsets) play fine with linux distributions. Yes, it is. I am seriously considering, when I get the right modem and know things are lined up and am starting production of the new system, to just order 20-30 of whatever I find. I'll just eat the cost then and make it up as I deploy them. I don't want to find the right modem, order 3-4, then find it not in production anymore. Again, I'd go with RS-232, but the expense is greater and, honestly, when they're going in someone else's office, so I have to be sure I can ssh in (which is a project in itself, considering different offices and different sysadmins), I have to be aware I will NOT have hands-on access. It sounds picky, but using an adaptor on a plug in such a case is just another thing that can go wrong and that I could spend hours trouble shooting to find some oaf jarred it and there's a loose connection. The fewer the connections and the simpler the system, the more time I get to spend ballroom dancing instead of patching software. (I'm a believer in setting up a system, then letting it do the work while I don't!) I have used a USB-to-RS-232 converter with success for a FOSS project to control an HD radio (http://halblog.com/hdradiocontroller.html), but an RS-232 modem is more expensive and the converters are expensive, too, putting the price per modem well over $50 each. I've ordered a Rosewill that looks good. It costs more than the one I have now. I'm going to try any drivers I can find for this one, but considering the price of the other one, with shipping, is $30, if the drivers don't work and the manufacturer doesn't help, I'm not going to knock myself out over this one. I can always use it on a Windows machine and use that system for testing (to have my other ones dial in). As it is, though, considering how cheap these are (this one and the replacement people have tried on Ubuntu), I think spending more than a few hours on this one might be a diminishing return. Still, if all it needs are drivers, then I'll be happy with it! O.k. I also think geting an USB modem to work should just be plug and play and no needing to mess with drivers at all. But it could worst: there are some embedded modems (those you can find in notebooks) that lack of any driver and they render completely useless. Yes
Re: Communicating with USB Modem
On Oct 9, 2010, at 10:17 PM, Phil Requirements wrote: On 2010-10-09 19:26:42 -0400, Hal Vaughan wrote: On Oct 9, 2010, at 6:49 PM, Camaleón wrote: [big snip] O.k. I also think geting an USB modem to work should just be plug and play and no needing to mess with drivers at all. But it could worst: there are some embedded modems (those you can find in notebooks) that lack of any driver and they render completely useless. Yes, that's true. And, unfortunately, I had information indicating that any USB modem would be similar to RS-232 in that it'd be plug-n-play. That's not so. And, while I'm looking at embedded computers (right now I'm waiting to see if the new Soekris Net-6501 will do well for me), at least the modems aren't embedded! And if I do use Soekris, they have a serial port -- you HAVE to use as a terminal during setup, so once I get the original image created and working, I can copy it to an image file and easily install it on flash cards and just insert it, without using the serial port on each one. And when they're deployed, I may be able to use it for an RS-232 modem, but I'm not committing to that yet. I had an external USB modem that was initially very gratifying, to use with a computer that had no serial port. I plugged it in, the system found it, and I was online in minutes. No external drivers. It was a very nice experience after having spent many hours trying to get drivers to work for WinModem cards. I would think, ideally, a good modem would be like that -- as long as it responds to AT commands properly, then it wouldn't need a driver. The one I used was US Robotics, but I can't lay my hand on it right this minute to say the product number. I saw some USR USB modems, but considering how I'll be buying a number of these, and footing the price myself (for various reasons, I can't really pass along this cost to my clients), if the $25 modem I ordered works, I'll be using it over the $45-$50 USR modems. It's pricing. Now if they're flaky, then, yes, I'll spend more, but if they work, I won't. I did have some problems with this modem over time. The main thing I remember is it would get moved from /dev/ttyUSB0 to /dev/ttyUSB1 occasionally. Or sometimes when the machine was booted the device wouldn't get established. I don't know if these were problems with the modem or my system. That's a USB issue. I ran into that when I was writing my code for controlling the Visteon HD radio (I have a copy of the project on my blog, but the main reason I did it was to turn it over so LinuxICE would have a radio controller). I had to write a routine that would first check the config file and look for the radio in its last place, then, if needed, send the same signal to every USB device and look for the response. The only issue is that I was worried the wake-up signal for the radio might crash some devices. I can't remember why I didn't just scan /proc/bus/usb/devices to get the info. I think I had to scan the /dev/ttySx devices. While the radio would USUALLY be at the same place it was last time, sometimes it would move. I wanted to point out that the modem I had initially seemed very good because of how easy installation was. But that's not the whole story. And that there are some plug and play USB modems out there. I think, for modems, plug and play means that it simply responds to the Hayes command set through a device in /dev. I'm not sure what you'd set the serial bus speed for in a program like minicom, though. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9db1b81f-0b95-4f4b-8aaf-c5939b913...@halblog.com
Can't Install Packages in Batch Mode
I have a bash script with the following in it: DEBIAN_FRONTEND=noninteractive apt-get install $install From what I've read, this should set the front end for debconf to non-interactive and should go on ahead with the install I've set up with a list of packages to install. Everything else in the script works, but as much as I've played with this (like changing DEBIAN_FRONTEND to DEBCONF_FRONTEND), I can't get it to work. I saw one example (from this list, 2-3 years ago) that put it all on one line, but that didn't work either. Do I have to set the environment variable before running the script? What do I need to do to be able to install packages? (There's only one question asked in all these packages and it's a config issue I handle separately anyway.) Thanks for any help with this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ccb30fe7-6deb-453d-bb9a-96cec7487...@halblog.com
Re: Can't Install Packages in Batch Mode
On Sep 26, 2010, at 5:08 PM, Hal Vaughan wrote: I have a bash script with the following in it: DEBIAN_FRONTEND=noninteractive apt-get install $install From what I've read, this should set the front end for debconf to non-interactive and should go on ahead with the install I've set up with a list of packages to install. Everything else in the script works, but as much as I've played with this (like changing DEBIAN_FRONTEND to DEBCONF_FRONTEND), I can't get it to work. I saw one example (from this list, 2-3 years ago) that put it all on one line, but that didn't work either. Do I have to set the environment variable before running the script? What do I need to do to be able to install packages? (There's only one question asked in all these packages and it's a config issue I handle separately anyway.) Thanks for any help with this! I have no idea what made the difference, but I re-typed the lines and re-tried and it worked. I had copied the lines (other than the $install variable) directly from an example. When it didn't work, I tried cutting and pasting from another example and it didn't work, but when I entered it by typing directly, it worked. Probably a typo that I couldn't catch with my learning disability or something. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/40856253-8b5f-4cd1-95bc-53249271b...@halblog.com
Re: Updating files in /etc Remotely (and automated)
On Sep 12, 2010, at 10:51 AM, Rob Owens wrote: On Sat, Sep 11, 2010 at 05:15:50PM -0400, Hal Vaughan wrote: I will be working with a server on the Internet that uses rsync and is running Debian. I will be setting up initial /etc/rsyncd.conf and /etc/rsyncd.secrets files on it. But along the way, whenever a new user is added, they'll need to be updated. I can use ssh on this system, but, of course, I don't want to allow root access. I'd like to be able to have these files updated automatically when I add a new user to another system. I could create new copies of the files locally, where the users are added and use scp to copy them to a directory on the server. But that's where there are problems. How can I chown the files to root, copy them to /etc, and chmod as needed for rsync to use them automatically? I don't see a way to do that without security issues. I need to somehow ssh in and do an su or run three commands as sudo (I need to mv the file, chown it, and chmod it). I am far from an expert in security, but I can see that if I have anything in place to make this easy, then anyone hacking my user account could easily mess up anything in the system. Is there some way I can set this up so I can update rsyncd.conf and rsyncd.secrets only automatically when I have the newer versions on my local system to be uploaded? When using ssh keys to log in, you can specify (in ~/.ssh/authorized_keys) a command which will automatically run when that key is used to log in. And that key will be useless to do anything else. Simply using that key to conenct to the remote server will run that command. The authorized_keys file would look something like this: command=/path/to/my/script ssh-rsa B3NzaC1yc2EAAA m...@myhost I see. That would make perfect sense and I see I can use -i to specify which key to use, so for normal situations, I just use ssh host, and when I want this done, I do ssh -i .ssh/special_key host instead. I thought I knew about authorized keys, but didn't know you could specify a command to be run in that file. You could use this to ssh into the remote server as root, or as a user with very specify sudo privileges that will allow your script to run. (The script would perform the file changes you need done, or simply rsync them from your local machine). But if I'm not running as root, from what I can see, no matter what I do with sudo, I still have to type in a password, don't I? using the authorized_keys file and specifying what can be done at login does a lot to help with security, but if I don't log in as root, no matter what I do, I'll still have to type in a password to use either su or sudo, right? Or is there a way around it? I was going through man pages, but it seems both require a password to be typed in no matter what. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9229c387-bb4b-4004-834a-3bea7fa77...@halblog.com
Re: Updating files in /etc Remotely (and automated)
On Sep 12, 2010, at 12:37 PM, Rob Owens wrote: On Sun, Sep 12, 2010 at 12:01:26PM -0400, Hal Vaughan wrote: On Sep 12, 2010, at 10:51 AM, Rob Owens wrote: On Sat, Sep 11, 2010 at 05:15:50PM -0400, Hal Vaughan wrote: I will be working with a server on the Internet that uses rsync and is running Debian. I will be setting up initial /etc/rsyncd.conf and /etc/rsyncd.secrets files on it. But along the way, whenever a new user is added, they'll need to be updated. I can use ssh on this system, but, of course, I don't want to allow root access. I'd like to be able to have these files updated automatically when I add a new user to another system. I could create new copies of the files locally, where the users are added and use scp to copy them to a directory on the server. But that's where there are problems. How can I chown the files to root, copy them to /etc, and chmod as needed for rsync to use them automatically? I don't see a way to do that without security issues. I need to somehow ssh in and do an su or run three commands as sudo (I need to mv the file, chown it, and chmod it). I am far from an expert in security, but I can see that if I have anything in place to make this easy, then anyone hacking my user account could easily mess up anything in the system. Is there some way I can set this up so I can update rsyncd.conf and rsyncd.secrets only automatically when I have the newer versions on my local system to be uploaded? When using ssh keys to log in, you can specify (in ~/.ssh/authorized_keys) a command which will automatically run when that key is used to log in. And that key will be useless to do anything else. Simply using that key to conenct to the remote server will run that command. The authorized_keys file would look something like this: command=/path/to/my/script ssh-rsa B3NzaC1yc2EAAA m...@myhost I see. That would make perfect sense and I see I can use -i to specify which key to use, so for normal situations, I just use ssh host, and when I want this done, I do ssh -i .ssh/special_key host instead. I thought I knew about authorized keys, but didn't know you could specify a command to be run in that file. You could use this to ssh into the remote server as root, or as a user with very specify sudo privileges that will allow your script to run. (The script would perform the file changes you need done, or simply rsync them from your local machine). But if I'm not running as root, from what I can see, no matter what I do with sudo, I still have to type in a password, don't I? using the authorized_keys file and specifying what can be done at login does a lot to help with security, but if I don't log in as root, no matter what I do, I'll still have to type in a password to use either su or sudo, right? Or is there a way around it? I was going through man pages, but it seems both require a password to be typed in no matter what. In /etc/sudoers, you can specify NOPASSWD, like this: someuser ALL=NOPASSWD: /path/to/some/command Then someuser can run the specified command as root without typing a password. My bad, in this case. I read the SUDO man page over a few times, but forgot to read SUDOERS man page. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/dd0542fe-3b19-4d26-a129-b03d831b0...@halblog.com
Re: Updating files in /etc Remotely (and automated)
On Sep 12, 2010, at 1:45 PM, Joe wrote: On 11/09/10 22:15, Hal Vaughan wrote: I will be working with a server on the Internet that uses rsync and is running Debian. I will be setting up initial /etc/rsyncd.conf and /etc/rsyncd.secrets files on it. But along the way, whenever a new user is added, they'll need to be updated. I can use ssh on this system, but, of course, I don't want to allow root access. I'd like to be able to have these files updated automatically when I add a new user to another system. I could create new copies of the files locally, where the users are added and use scp to copy them to a directory on the server. But that's where there are problems. How can I chown the files to root, copy them to /etc, and chmod as needed for rsync to use them automatically? I don't see a way to do that without security issues. I need to somehow ssh in and do an su or run three commands as sudo (I need to mv the file, chown it, and chmod it). I am far from an expert in security, but I can see that if I have anything in place to make this easy, then anyone hacking my user account could easily mess up anything in the system. Is there some way I can set this up so I can update rsyncd.conf and rsyncd.secrets only automatically when I have the newer versions on my local system to be uploaded? Thanks for any ideas! Hal How quickly do you need the updates? Cron will run scripts as root, and can run your script as often as you can stand the overhead. You just need to get the files there in a safe way. I had completely overlooked that idea and someone sent it to me privately a little while ago. While I like what Rob Owens suggested, I'm leaning toward this. I think it's possible that I could send up the minimum information in a file and have the cron job be a Perl script that takes that info and builds the rsyncd.conf and rsyncd.secrets files from there, which reduces the possibility of a rogue file being copied over somehow. Still, none of the ideas is perfect, but putting together the conf files on the site, as opposed to sending them directly, has certain merits. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/c04810b6-f31e-45a7-ad91-c8d5fe13f...@halblog.com
Re: Updating files in /etc Remotely (and automated)
On Sep 12, 2010, at 12:37 PM, Rob Owens wrote: ... When using ssh keys to log in, you can specify (in ~/.ssh/authorized_keys) a command which will automatically run when that key is used to log in. And that key will be useless to do anything else. Simply using that key to conenct to the remote server will run that command. The authorized_keys file would look something like this: command=/path/to/my/script ssh-rsa B3NzaC1yc2EAAA m...@myhost I see. That would make perfect sense and I see I can use -i to specify which key to use, so for normal situations, I just use ssh host, and when I want this done, I do ssh -i .ssh/special_key host instead. I thought I knew about authorized keys, but didn't know you could specify a command to be run in that file. You could use this to ssh into the remote server as root, or as a user with very specify sudo privileges that will allow your script to run. (The script would perform the file changes you need done, or simply rsync them from your local machine). But if I'm not running as root, from what I can see, no matter what I do with sudo, I still have to type in a password, don't I? using the authorized_keys file and specifying what can be done at login does a lot to help with security, but if I don't log in as root, no matter what I do, I'll still have to type in a password to use either su or sudo, right? Or is there a way around it? I was going through man pages, but it seems both require a password to be typed in no matter what. In /etc/sudoers, you can specify NOPASSWD, like this: someuser ALL=NOPASSWD: /path/to/some/command Then someuser can run the specified command as root without typing a password. When I tested this with some simple scripts, I find if I create a batch file that runs a few commands, like chown root:root filename that those commands, which would normally need the sudo command don't need it. Is this because of the (usually) 5 minute time limit sudo uses? Can I trust this on all systems, or is there anything that could prevent this behavior? In other words, if I include, in the script, commands that also need sudo, am I right that I can count on them executing without further need of verification? Thanks for anyone who can include more info on this! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/0ea47e68-49ed-4f32-a91b-4379e27e4...@halblog.com
Updating files in /etc Remotely (and automated)
I will be working with a server on the Internet that uses rsync and is running Debian. I will be setting up initial /etc/rsyncd.conf and /etc/rsyncd.secrets files on it. But along the way, whenever a new user is added, they'll need to be updated. I can use ssh on this system, but, of course, I don't want to allow root access. I'd like to be able to have these files updated automatically when I add a new user to another system. I could create new copies of the files locally, where the users are added and use scp to copy them to a directory on the server. But that's where there are problems. How can I chown the files to root, copy them to /etc, and chmod as needed for rsync to use them automatically? I don't see a way to do that without security issues. I need to somehow ssh in and do an su or run three commands as sudo (I need to mv the file, chown it, and chmod it). I am far from an expert in security, but I can see that if I have anything in place to make this easy, then anyone hacking my user account could easily mess up anything in the system. Is there some way I can set this up so I can update rsyncd.conf and rsyncd.secrets only automatically when I have the newer versions on my local system to be uploaded? Thanks for any ideas! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ef953506-5924-409c-b345-b43c694dd...@halblog.com
Re: SSH Rsync issues
On Sep 10, 2010, at 8:14 AM, Rob Owens wrote: On Thu, Sep 09, 2010 at 06:01:55PM -0400, Hal Vaughan wrote: In short: I have ssh set up on two systems so I can ssh from one to the other. My id_rsa.pub in ~/.ssh on my system is copied into ~/.ssh/authorized_keys on the remote system. I can ssh from local to remote with no issue and it's configured so authentication does not use passwords, but uses the RSA ID. This works perfectly. ssh remote gets me logged in immediately. I can rsync to the other machine. Using rsync localfile tnet-web::threshNet-Public works fine and the file is transferred. BUT when I try to use rsync over ssh, it will NOT work Any suggestions? I think you are mixing/confusing the 2 rsync methods. One is the rsyncd daemon. To rsync to an rsyncd daemon, you use two colons after the hostname, like this rsync localfile tnet-web::threshNet-Public This tranfers everything in the clear. Last I checked, there was no built-in method to transfer over ssh using the daemon. You could set up an SSL or SSH tunnel to do that. The other method of using rsync is with a single colon, like this rsync localfile tnet-web:threshNet-Public This automatically uses ssh for transfer, but it requires a few things. 1) your user must have an account on tnet-web 2) threshNet-Public is a folder inside your user's home directory on tnet-web 3) it does not require an rsyncd.conf file, and doesn't look at it even if you have one. Hope that helps. Yes, that clarifies it. I was not aware of the one colon vs. the two colon situation at all and it makes a lot of sense. But that still leaves the --rsh=ssh option as a question. From reading the man pages, I was thinking that would make rsync use ssh, but nothing I've done made it work. I've decided to encrypt the files before syncing them so they won't be sent in the clear and to use passwords on rsync. Since a number of different client systems will use this setup, I don't want them all having ssh keys or access to the system through ssh in case of a break-in. Thank you. You've cleared up a LOT for me. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2f4c87e7-585a-4704-bbd3-33bfbb945...@halblog.com
Re: SSH Rsync issues
On Sep 10, 2010, at 1:36 PM, Rob Owens wrote: On Fri, Sep 10, 2010 at 10:57:56AM -0400, Hal Vaughan wrote: On Sep 10, 2010, at 8:14 AM, Rob Owens wrote: On Thu, Sep 09, 2010 at 06:01:55PM -0400, Hal Vaughan wrote: In short: I have ssh set up on two systems so I can ssh from one to the other. My id_rsa.pub in ~/.ssh on my system is copied into ~/.ssh/authorized_keys on the remote system. I can ssh from local to remote with no issue and it's configured so authentication does not use passwords, but uses the RSA ID. This works perfectly. ssh remote gets me logged in immediately. I can rsync to the other machine. Using rsync localfile tnet-web::threshNet-Public works fine and the file is transferred. BUT when I try to use rsync over ssh, it will NOT work Any suggestions? I think you are mixing/confusing the 2 rsync methods. One is the rsyncd daemon. To rsync to an rsyncd daemon, you use two colons after the hostname, like this rsync localfile tnet-web::threshNet-Public This tranfers everything in the clear. Last I checked, there was no built-in method to transfer over ssh using the daemon. You could set up an SSL or SSH tunnel to do that. The other method of using rsync is with a single colon, like this rsync localfile tnet-web:threshNet-Public This automatically uses ssh for transfer, but it requires a few things. 1) your user must have an account on tnet-web 2) threshNet-Public is a folder inside your user's home directory on tnet-web 3) it does not require an rsyncd.conf file, and doesn't look at it even if you have one. Hope that helps. Yes, that clarifies it. I was not aware of the one colon vs. the two colon situation at all and it makes a lot of sense. But that still leaves the --rsh=ssh option as a question. From reading the man pages, I was thinking that would make rsync use ssh, but nothing I've done made it work. I've decided to encrypt the files before syncing them so they won't be sent in the clear and to use passwords on rsync. Since a number of different client systems will use this setup, I don't want them all having ssh keys or access to the system through ssh in case of a break-in. Thank you. You've cleared up a LOT for me. On my Lenny system, man rsync states the following in the CONNECTING TO AN RSYNC DAEMON section: you must not specify the --rsh (-e) option I had never realized man pages differed so much. I even found differences on the man pages on my iMac. (Can I say iMac on this list or is it a dirty word here?) There's also a section USING RSYNC-DAEMON FEATURES VIA A REMOTE-SHELL CONNECTION that you might find useful. It mentions some ways of encrypting the rsync daemon transfer. It also shows the following command, which seems to ignore the you must not specify... rule above: rsync -av --rsh=ssh host::module /dest But if you read carefully it seems to say that this requires a user account on the host machine. I've never tried that, so I can't help you much there. That's where I got confusing -- they give an example, the very one you point out, and I tried it, but could not get it to work, even with a user who had an account on the receiving system. The remote system, though, was on my web hosting service and it was later I realized they have a different version of rsync, which could result in compatibility issues or that version not doing what others can do. I did try it, though, exactly like they described, and with a user who had a shell account on the remote system. You should probably google rsyncd encryption and see what you can find. For the single-colon rsync, you don't need to specify --rsh=ssh. It is the default. Yeah, but I don't want to set up user accounts on the host. For one thing, on my web hosting site, Westhost doesn't provide an easy way to add users, so I can't just add another easily. Everything in my system is automated so I can add a new client/user with a single command. It's a pain to have it all set up here then have to go to the web control panel on the website to add a user. When it's not automated, it's easy to forget a step of the process. I've decided I'm going to encrypt the files locally, then send them up using rsync to an account that requires a password and the other system will download them THEN unencrypt them, so the files will be encrypted when sent over the Internet and stored there and only clear when they're on a local system. Thanks for the info on this. I'm going to give up, in this case, on the whole ssh thing with rsync. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2a5a5e5e-f874-4e47-b933-591882c8f...@halblog.com
Re: SSH Rsync issues
On Sep 10, 2010, at 2:49 PM, Rob Owens wrote: On Fri, Sep 10, 2010 at 02:18:42PM -0400, Hal Vaughan wrote: On Sep 10, 2010, at 1:36 PM, Rob Owens wrote: You should probably google rsyncd encryption and see what you can find. For the single-colon rsync, you don't need to specify --rsh=ssh. It is the default. Yeah, but I don't want to set up user accounts on the host. For one thing, on my web hosting site, Westhost doesn't provide an easy way to add users, so I can't just add another easily. Everything in my system is automated so I can add a new client/user with a single command. It's a pain to have it all set up here then have to go to the web control panel on the website to add a user. When it's not automated, it's easy to forget a step of the process. I've decided I'm going to encrypt the files locally, then send them up using rsync to an account that requires a password and the other system will download them THEN unencrypt them, so the files will be encrypted when sent over the Internet and stored there and only clear when they're on a local system. Just be careful. I think I recall reading that the rsyncd user/password is sent either cleartext or with not-too-difficult-to-crack encryption. Here is a fairly old writeup on using rsyncd with stunnel. Maybe it will be helpful. http://www.netbits.us/docs/stunnel_rsync.html Thanks for the tip and the info on stunnel! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/bebaf511-39e2-4de3-b6c3-4796cb142...@halblog.com
SSH Rsync issues
In short: I have ssh set up on two systems so I can ssh from one to the other. My id_rsa.pub in ~/.ssh on my system is copied into ~/.ssh/authorized_keys on the remote system. I can ssh from local to remote with no issue and it's configured so authentication does not use passwords, but uses the RSA ID. This works perfectly. ssh remote gets me logged in immediately. I can rsync to the other machine. Using rsync localfile tnet-web::threshNet-Public works fine and the file is transferred. BUT when I try to use rsync over ssh, it will NOT work. A bit more of an explanation: The problem is that when I try rsync localfile - --rsh=ssh ... or rsync localfile - -essh -l myname ... (Yes, same stuff as before with the ...) Then I get errors: using the -e option gives me this: rsync: -rsh=ssh: unknown option _exit_cleanup(code=1, file=/SourceCache/rsync/rsync-40/rsync/main.c, line=1333): entered rsync error: syntax or usage error (code 1) at /SourceCache/rsync/rsync-40/rsync/main.c(1333) [client=2.6.9] _exit_cleanup(code=1, file=/SourceCache/rsync/rsync-40/rsync/main.c, line=1333): about to call exit(1) using --rsh gives me this: cmd=ssh machine=tnet-web user= path=threshNet-Public cmd[0]=ssh cmd[1]=tnet-web cmd[2]=rsync cmd[3]=--server cmd[4]=--daemon cmd[5]=. opening connection using ssh tnet-web rsync --server --daemon . rsync: connection unexpectedly closed (0 bytes received so far) [sender] _exit_cleanup(code=12, file=/SourceCache/rsync/rsync-40/rsync/io.c, line=452): entered rsync error: error in rsync protocol data stream (code 12) at /SourceCache/rsync/rsync-40/rsync/io.c(452) [sender=2.6.9] _exit_cleanup(code=12, file=/SourceCache/rsync/rsync-40/rsync/io.c, line=452): about to call exit(12) I have config files for rsyncd.conf and sshd_conf included below my signature. Both programs work separately, but if I try rsync over ssh, it simply does not work. Any suggestions? Thanks! Hal -- rsyncd.conf: #Globals: log file = /var/log/rsyncd.log secrets file = /etc/rsyncd.secrets [threshNet] path = /threshNet comment = Full threshNet system backup read only = no #auth users = hal [threshNet-Hal] comment = threshNet test comm area path = /threshNet/Hal read only = no auth users = hal [threshNet-TNTest] comment = threshNet communications area for TNTest path = /threshNet/TNTest read only = no auth users = TNTest [threshNet-Public] comment = threshNet public comm directory for testing path = threshNet/Public read only = no sshd_conf: - # Package generated configuration file # See the sshd(8) manpage for details # What ports, IPs and protocols we listen for Port 22 # Use these options to restrict which interfaces/protocols sshd will bind to #ListenAddress :: #ListenAddress 0.0.0.0 Protocol 2 # HostKeys for protocol version 2
Re: SSH Rsync issues
On Sep 9, 2010, at 6:46 PM, Kevin Ross wrote: On 09/09/2010 03:01 PM, Hal Vaughan wrote: I can rsync to the other machine. Using rsync localfile tnet-web::threshNet-Public works fine and the file is transferred. BUT when I try to use rsync over ssh, it will NOT work. According to the man page, your first example should automatically use ssh. It does for me. Do you mean the first example in my first paragraph, where there's no reference to ssh or the one when I list examples, where I use --rsh =ssh? How can I verify rsync is using ssh? I found a few comments in man pages that left me confused, but I read man pages on Linux, on OS X, and on the web and there were a few references in some to, in future versions, so it's hard to be sure, unless it's explicitly stated. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/f44e19b6-6213-4ffe-8bc4-373cad26b...@halblog.com
Re: DHClient Exit Scripts
On Aug 15, 2010, at 5:41 PM, Hal Vaughan wrote: I've been reading the man pages for dhclient, but I'm stuck with one key question I've missed. I have an exit script set up to notify me of the IP address of a particular system (on a LAN) whenever the IP address changes. For now I'm testing, but in the future it'll be running where I have no control over DHCP or any servers. This system will be just plugged into the power outlet and into a CAT5 cable hooked to a hub or switch, so it has to get a dynamic IP address from the DHCP and it has to be able to alert me to what the new IP address is. I know whenever dhclient runs, it runs the exit scripts. The one issue I'm not clear of is whether the exit scripts will be run any time the IP address changes, like when a lease expires. I know it'll run at reboot or if networking is restarted, but will the exit scripts be run when the IP address is changed by the DHCP? If it's not, is there any program that runs under those conditions? I had different answers from different people and finally wrote to Ted Lemon, the author of dhclient to clarify. Here is my note to him with his response: - On Aug 16, 2010, at 9:57 AM, Ted Lemon wrote: On Aug 16, 2010, at 12:05 AM, Hal Vaughan wrote: If it is running as a daemon, will it run dhclient-script whenever the IP on an interface is changed, such as when the lease runs out or when the DNS changes the address for some other reason? dhclient-script runs whenever a transaction completes, whether it's an INIT, an INIT-REBOOT, a RENEW, or a REBIND. It also runs when a transaction times out, or when the client *fails* to renew a lease. Its purpose in life is to take whatever happened on a protocol level and translate it into configuration actions. At renewal, it's responsible for *noticing* that things have changed, like the DNS server, and updating the resolv.conf file. - A few notes on this: 1) On Debian there is a directory, /etc/dhcp3/dhclient-exit-hooks.d instead of one exit script. Any scripts in that directory will be run when dhclient-script runs, which, as Ted clarified, is on any event involving the interface, instead of just running one script (which would force one to put commands at the end of that one script for each additional script one needs to run). 2) It will only run bash scripts. I tried putting a Perl script in that directory and it wouldn't work, so I had to put a bash script in that directory to run my Perl script. 3) The packages I found that work with DynDNS or similar services mostly ran Perl or Python scripts and were daemons, meaning that using one would mean keeping an instance of Perl or Python in memory, in addition to the needed program. In my case, I'm working on an embedded system, so that's too much memory to use. Since dhclient is used by default on most systems, there's no need for an additional daemon, just a script that dhclient-script can run to update a dynamic DNS service. (Why run two daemons when you need only one?) I may modify one of the scripts in the Debian packages and send the info back to the author in case he wants to let it work with dhclient as an option. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/84960162-4435-43b2-a07b-3361f8bda...@halblog.com
Re: DHClient Exit Scripts
On Aug 17, 2010, at 12:30 PM, Boyd Stephen Smith Jr. wrote:
In 84960162-4435-43b2-a07b-3361f8bda...@halblog.com, Hal Vaughan wrote:
2) It will only run bash scripts. I tried putting a Perl script in that
directory and it wouldn't work, so I had to put a bash script in that
directory to run my Perl script.
This seems wrong. Did your perl script have the correct she-bang line as
the first line of the file: #!/usr/bin/perl ? Did it have execute
permissions?
I agree, it seems wrong, but that was my experience. I did use the correct
she-bang, I even copied that Perl script, as it was, into ~/bin, without
changing it. However, the permissions for the scripts in
/etc/dhcp3/dhclient-exit-hook.d are all -rw-r--r-- and owned by root. I had
checked to make sure my Perl script matched in ownership and permissions, still
no go. Of course, when I put it in my ~/bin directory, I changed the ownership
to my account and changed the permissions so it was a normal executable. And
my bash script I replaced it with in /etc/dhcp3/dhclient-exit-hook.d matched in
ownership and permissions to all the other files in that directory.
Like many parts of Debian, I'd expect this to be using run-parts or
equivalent, which simply makes a C/system call to exec(). On Linux, exec()
handles ELF executables with the +x bit and text files with a she-bang line
and the +x bit.
But dhclient-script, which is what is run on any event involving the interface,
is a bash script, not a C program. I read that it's 255 (or was it 254) lines
long when I scanned it in less. There's an exit_with_hooks() function it calls
at the end. First it looks for the script /etc/dhcp3/dhclient-exit-hooks and
if it exists, it runs it with the routine run_hook(). Then it looks for the
directory /etc/dhcp3/dhclient-exit-hooks.d (which, apparently, from comments,
is a Debian only thing) and calls run_hookdir() to run the scrips in that
directory. For each script, it calls run_hook() to run it. Here's the routine
run_hook():
run_hook() {
local script=$1
local exit_status
shift # discard the first argument, then the rest are the script's
if [ -f $script ]; then
. $script $@
fi
if [ -n $exit_status ] [ $exit_status -ne 0 ]; then
logger -p daemon.err $script returned non-zero exit status
$exit_status
save_exit_status=$exit_status
fi
return $exit_status
}
I'm not an expert on bash scripting, but if I remember right, the dot command
is the same as source where, rather than running a script, it run the
commands in the script. I could be wrong about that, but if that's the case,
then it explains why it executes a bash script and not a Perl script.
Hal
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/c6ae9aba-68c4-4b94-8bb8-8c77ed2a9...@halblog.com
DHClient Exit Scripts
I've been reading the man pages for dhclient, but I'm stuck with one key question I've missed. I have an exit script set up to notify me of the IP address of a particular system (on a LAN) whenever the IP address changes. For now I'm testing, but in the future it'll be running where I have no control over DHCP or any servers. This system will be just plugged into the power outlet and into a CAT5 cable hooked to a hub or switch, so it has to get a dynamic IP address from the DHCP and it has to be able to alert me to what the new IP address is. I know whenever dhclient runs, it runs the exit scripts. The one issue I'm not clear of is whether the exit scripts will be run any time the IP address changes, like when a lease expires. I know it'll run at reboot or if networking is restarted, but will the exit scripts be run when the IP address is changed by the DHCP? If it's not, is there any program that runs under those conditions? Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1af890a5-5bf9-46c9-8c4b-b709170dd...@halblog.com
Re: DHClient Exit Scripts
On Aug 15, 2010, at 6:37 PM, Boyd Stephen Smith Jr. wrote: In 1af890a5-5bf9-46c9-8c4b-b709170dd...@halblog.com, Hal Vaughan wrote: I've been reading the man pages for dhclient, but I'm stuck with one key question I've missed. I have an exit script set up to notify me of the IP address of a particular system (on a LAN) whenever the IP address changes. I know whenever dhclient runs, it runs the exit scripts. The one issue I'm not clear of is whether the exit scripts will be run any time the IP address changes, like when a lease expires. I know it'll run at reboot or if networking is restarted, but will the exit scripts be run when the IP address is changed by the DHCP? If it's not, is there any program that runs under those conditions? I can't answer the specific question, but I knows that there are DynDNS clients in Debian that are capable of notifying a remote system whenever the local system's IP changes. One of these clients should be able to speak a dialect you can process. I had always thought those were for external IP addresses, not for inside a LAN. I'll check on that. Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/434131f9-b1ad-46bf-901c-ad3be8f07...@halblog.com
Re: Embedded System Recommendations
On Jul 28, 2010, at 4:58 PM, Volkan YAZICI wrote: On Wed, 28 Jul 2010, Hal Vaughan h...@halblog.com writes: I'm working with something that will go in people's offices. I was seriously considering using some Soekris boxes (http://soekris.com). The problem is the upper limit of RAM is 512 MB. I'd like to get something a little faster than their systems and with more memory, but the memory would take priority. Did you check EPIA Mini-ITX boards? My big issue with those (and some others) is that if I want to block off video and keyboard ports (which, these days, means USB, too), I have to physically alter the case or the board or do something like use glue to seal the connector. I want to make this as much appliance like as possible so there's no temptation to think of it like a computer or to try to log in or connect to it normally. The Soekris boxes have a serial port for communications and, other than something like ssh, that's it for logging in. But thanks for the suggestion! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/dbd08524-0d21-440c-b3e4-d90351e16...@halblog.com
Re: Embedded System Recommendations
On Jul 29, 2010, at 2:06 PM, Perry E. Metzger wrote: On Wed, 28 Jul 2010 15:41:11 -0400 Hal Vaughan h...@halblog.com wrote: I'm working with something that will go in people's offices. I was seriously considering using some Soekris boxes (http://soekris.com). The problem is the upper limit of RAM is 512 MB. I'd like to get something a little faster than their systems and with more memory, but the memory would take priority. I don't want to go with a normal system because I don't want video or keyboard attached. With Soekris, I need a null modem serial cable to connect when I'm setting it up and installing Debian. (Then, of course, ssh works fine.) I want headless because I want it to feel like an appliance to the people I'm working with. I don't want them trying to access it directly (it'll have a web interface). Does anyone use Debian on a box like this (headless) that they can recommend? If you're willing to leave the world of x86, there are lots of cool things out there like the Beagleboard, Shivaplug/Guruplug, etc. Okay, that's a start. I have no problem with leaving x86, just as long as I can easily get Debian running on it -- and can easily replicate that when I need more. Any other names that you know work with Debian? Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/b5e4137b-ea6b-4097-b27d-f5ed6eb3e...@halblog.com
Embedded System Recommendations
I'm working with something that will go in people's offices. I was seriously considering using some Soekris boxes (http://soekris.com). The problem is the upper limit of RAM is 512 MB. I'd like to get something a little faster than their systems and with more memory, but the memory would take priority. I don't want to go with a normal system because I don't want video or keyboard attached. With Soekris, I need a null modem serial cable to connect when I'm setting it up and installing Debian. (Then, of course, ssh works fine.) I want headless because I want it to feel like an appliance to the people I'm working with. I don't want them trying to access it directly (it'll have a web interface). Does anyone use Debian on a box like this (headless) that they can recommend? Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/dfc7789f-5284-4535-89cf-c9d71b158...@halblog.com
Re: Esperanto
On Apr 28, 2010, at 1:22 AM, Robert Holtzman wrote: On Tue, 27 Apr 2010, Celejar wrote: On Tue, 27 Apr 2010 15:35:48 -0700 (MST) Robert Holtzman hol...@cox.net wrote: ... Isn't this list moderated? If so, why hasn't this imbecile been dumped? This list is not moderated. Too bad. I don't normally like list moderation but I would sure make an exception here. BTW he emailed me direct telling me to f***k myself and calling me a piece of s**t. He didn't have the guts to put it on the list. The quality of trolls sure has slipped lately. Think of it this way: We tell the rest of the world who we are by our actions. His trolling told us a lot more about him than he wanted us to know (like how board and unimaginative he is) and his response to you, and lack of courage to put it on the list, tells us even more, and none of it speaks well of him. He's told a few hundred people (or more?) a lot about his insecurities and lack of character. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9656f9c1-a204-46ad-8d6a-8914d70d7...@halblog.com
Re: BonziBuddy
On Oct 20, 2009, at 12:57 AM, Bonzi Buddy wrote: hi my friend wants me to install bonzibuddy 4 him but i cannot find it where can i get ithow should i install plz help this is a emergency k thx bai How is the need to install Bonzi Buddy an emergency? It's not something that one really needs and, really, it's more adware and spyware than a useful program. Oh, and there's a cool feature in most browsers now: It's called spellcheck. If you use it, then people, intelligent people, can understand what you write. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: no fsck?
On Jul 21, 2009, at 12:39 AM, jeremy jozwik wrote: new to debian, trying to restore some files $ ./fsck.vfat -rv /~~~ bash: ./fsck.vfat: No such file or directory please help! You've got fsck, you don't have fsck to work on vfat file systems, which are under the MS-DOG category. Try, as root, or using sudo, to run this: apt-get install dosfstools or, if you use aptitude: aptitude install dosfstools And that will add the necessary tools to work with those file systems. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 6, 2009, at 2:30 AM, Dotan Cohen wrote: I like to offend midgets. Can we do that now? Just a few small jokes? I'd prefer if you do a Jew joke! I'll do my best not to get offended. Anybody else volunteer for humiliation? I'm a Trekkie. Unfortunately, I don't live in my Mother's basement and I did kiss a few girls before I was 30. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 5, 2009, at 1:50 AM, Rustam wrote: On Sun, 2009-07-05 at 00:33 +0300, Dotan Cohen wrote: ... [snip] ... haha :) well, i know some might think this thread is offended, some might just delete. the truth is this is OT. but despite that, let me post my first reply to this: It doesn't work that way, I've been trying for ages! ... jaun...@laptop:~$ beg bash: beg: command not found rus...@ficus:~$ whatis beg beg: nothing appropriate. haha :D it doesn't work too. can we make linux more human ? Then it'd be called a Mac. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 5, 2009, at 2:32 AM, Dotan Cohen wrote: haha :D it doesn't work too. can we make linux more human ? Then it'd be called a Mac. Fanboy! Human is the Ubuntu colour/icon scheme! Fanboy! Ubuntu is Debian Sid Lite pretending to be some other OS and GUI. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 5, 2009, at 1:55 PM, Dotan Cohen wrote: It's too bad that some people are so thin-skinned that a word like boobies upsets them. It's not like we're talking about raping someone, for instance. Now I'm really offended. The thread stops here. And now we know the thread will go on at least 3 more days. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 4, 2009, at 3:28 PM, Cybe R. Wizard wrote: On Sat, 4 Jul 2009 21:44:51 +1200 Chris Bannister mockingb...@earthlight.co.nz wrote: On Sun, Jun 28, 2009 at 08:47:09AM -0500, Nate Bargmann wrote: * Comcast Master Account elizabethmusb...@comcast.net [2009 Jun 27 15:05 -0500]: hi i want boobies so i try grep some boobies but command not work how i get boobies thx I guess that if you can't grep any boobies, then the following commands will be difficult to invoke: unzip, strip, touch, finger, grep, mount, fsck, more, yes, fsck, fsck, fsck, umount, sleep $ man woman No manual entry for woman $ locate woman $ $ sudo mount /woman mount: can't find /woman in /etc/fstab or /etc/mtab (you bet I'm re-naming one of my storage partitions now!) $ sudo apt-get install woman Reading package lists... Done Building dependency tree Reading state information... Done E: Couldn't find package woman $ Oh, oh. Out of luck, I'd guess. ...unless there might still be some women somewhere on Usenet. Isn't that what alt.binaries.* is for? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 4, 2009, at 5:07 PM, Lisi Reisz wrote: On Saturday 04 July 2009 20:33:14 Dotan Cohen wrote: $ sudo apt-get install woman Reading package lists... Done Building dependency tree Reading state information... Done E: Couldn't find package woman $ Oh, oh. Out of luck, I'd guess. ...unless there might still be some women somewhere on Usenet. alias woman locate; talk; date; uptime; gawk; head; clean; sleep And then they wonder why some women find IT a hostile area Seriously, there are female readers of this list. Could we at least pretend that the rest of you think we are human beings, not blow-up toys for men to grope? I thought a large part of the joke was self-reflexive, making fun of the stereotypical view of how geeks and nerds supposedly don't understand romantic relationships, in other words, that it was as much people making fun of the stereotypes they could fall into as it was anything else. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jul 4, 2009, at 8:43 PM, Nate Bargmann wrote: * Lisi Reisz lisi.re...@gmail.com [2009 Jul 04 16:10 -0500]: On Saturday 04 July 2009 20:33:14 Dotan Cohen wrote: $ sudo apt-get install woman Reading package lists... Done Building dependency tree Reading state information... Done E: Couldn't find package woman $ Oh, oh. Out of luck, I'd guess. ...unless there might still be some women somewhere on Usenet. alias woman locate; talk; date; uptime; gawk; head; clean; sleep And then they wonder why some women find IT a hostile area Seriously, there are female readers of this list. Could we at least pretend that the rest of you think we are human beings, not blow-up toys for men to grope? Could it be that we guys are poking fun at ourselves? Is that not possible? 'zactly what I thought. In other words, please can we lay this rather offensive thread to rest? I guess that since I'm an insensitive clod of a male that I can't see the offensive nature of this thread. You do realize that your plea to end this thread only caused it to be continued longer. Ironic, no? Streisand effect, especially known to be prevalent on Debian-User. Or, as Hal's Law states: The surest way to make sure a thread continues ad nauseam on Debian-User is to give even the slightest suggestion it may be off-topic. Such an action will guarantee the thread will continue for at least another 72 hours. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how do i grep boobies
On Jun 27, 2009, at 4:53 PM, Scarletdown wrote: On Sat, Jun 27, 2009 at 1:34 PM, John Musbach johnmusba...@gmail.com wrote: I apologize, I had a virus on my computer and it was doing all kinds of weird stuff. Sorry! On Sat, Jun 27, 2009 at 4:01 PM, Comcast Master Accountelizabethmusb...@comcast.net wrote: hi i want boobies so i try grep some boobies but command not work how i get boobies thx Even so, I still have to say Get a grep on yourself. :) Is a text processing tool really the most appropriate tool for this job? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Goodbye debian
On Jun 25, 2009, at 4:37 PM, Roger Leigh wrote: On Thu, Jun 25, 2009 at 08:30:33AM +0200, Javier Barroso wrote: Hi, 2009/6/25 明覺 shi.min...@gmail.com: I do not have time to read your replies about another discussion anymore for they are useless, and I do not feel happy with all you debian guys, so I leave this mailing list, also debian has no attraction to me anymore, I will stop using it from now. As I have decalared, I will build my own OS and applications by a Only One Programming Lanuguage way. Good bye! :) Your original question was about installing debian without perl, it is not possible now. If you want to do it, you would have asked in debian-devel, because debian users can't do anything with this issue. Given the ghastliness of maintaining Perl code, rewriting it in C++ does have a certain attraction. However, one first has to understand what the code *does*, and a few thousand lines of uncommented regular expressions are a fairly impenetrable mess! I was teaching myself 6502 Assembler in college while I was taking a course on the Vax 11/780. My instructor had a rule in Assembler: EVERY line of code HAD to have a comment. Most people had lame comments, like, Transfer from register to memory. I had learned from reading Nibble Magazine, which published programs for the Apple ] [, ][+, and //e computers so if I had a task to do, I'd put in comments like, Take the given value and add the offset we need, then transfer it to use as an index. That might be split over 2-3 lines of code, but the teacher commented that he liked my comments because it made it clear what the program was doing. That was enough to start a lifelong habit. I don't comment every line now, but I use block comments for sections and subroutines and that habit has made my life MUCH easier. Also, by the time I started using Java, I had learned that many times when I write something, I'm revisiting later and have forgotten it, so making sure my Javadoc comments were clear enough that I wouldn't have to dig up the code just re-enforced that. I'm still surprised at the number of uncommented regex stuff I see in Perl. If I'm using a regex that has ANY degree of complexity (other than pulling out, say, one string with a wildcard or two), then I comment that line of code specifically. I've been thankful, many times, for the habit that teacher created with his requirements. I might add that it's not just Perl, it really depends on who wrote the code. Uncommented code can be a nightmare in ANY language! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 1:21 AM, Cowley Harris wrote: This guy asked a relatively simple question which I'm paraphrasing here as can you run Debian without perl or python, the answer is pretty much no. He gave his reasons for the question and his opinion on the answers he was given. He's also started probably the most interesting thread on this list for a while. I think the interest has come from people responding to him, not from anything he's said. As I've read it, he's not attacked anyone personally even when disagreeing with people and yet he is being personally attacked for his opinions. Well, here is my opinion, if you feel threatened enough by someone who disagrees with you, that you must insult them, it's a sign of a weak logic or a weak mind. It's the reaction typical of a zealot, fool or troll and not the response of someone with some useful knowledge to share. I don't see anyone being threatened by him or acting like they're threatened. I know I've stated my thoughts on this as I've seen some patterns emerge, such as his statements getting more and more extreme and as many of them show more and more of a claim to knowing something it's clear he does not know. But I haven't seen anyone insult him yet -- but then I haven't read many of the overnight posts yet. I don't agree with him that a one programming language system would be the right way to do things or that it would lead to a bright future of our free software world. but I'm not going to insult him for his belief. In fact it might be a good thing that he tries this endeavor. “Reasonable people adapt themselves to the world. Unreasonable people attempt to adapt the world to themselves. All progress, therefore, depends on unreasonable people.” GB Shaw. Yes, that's a cute quotation. I can also cite Zen quotations or other sources about the wisdom of knowing when to fight and when not to, or when to push a big rock out of the way or when it's better to walk around that rock. The benefits of multiple languages over a 1 language system (1LS) is that it gives you the ability to program at the appropriate level of the problem space. The studies show that the higher level you program the more instructions you have per line, and yet the bugs per line stay about the same. The studies also show that the amount of LOC produced by programmers of the same skill level is about the same, whether they use assembly, c or java, and yet the amount of instructions per LOC increase with level of the language. More productivity in the same time span is a major advantage. Definitely. There are definite advantages to a higher level language and we are well past the days when we gained any significant advantage by having a programmer write in Assembler to save memory space and CPU time instead of having the same system done in a higher level language, whether it's an interpreted one or compiled one. On a personal note, I think that programmers that use different languages can communicate in meta-programming terms, a hash-table is a hash-table whether you call it a hash or a dictionary. The benefits of a 1LS would be small in comparison to the benefits of plural system. Which is what pretty much everyone but the OP has said. I think we all agree on that. I've found times where I may be talking with someone who knows one language but does not know the one I'm working in that I'm using a modified hash table or I'll say, It's something like a hash table, but it's my own class, so there's some extra functions. That generally does well enough. I've never had a problem talking with another programmer who is not familiar with the language I'm using where I couldn't communicate what I was doing easily. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 1:53 AM, 明覺 wrote: On Wed, Jun 24, 2009 at 12:32 AM, Micha Feiginmi...@post.tau.ac.il wrote: On Mon, 22 Jun 2009 12:21:22 -0400 Hal Vaughan h...@halblog.com wrote: ... I'm being blunt, but, honestly, I run a business on custom software I've written and I can do it because I learned from those who knew more than I did. If I refused to learn from people on this and other lists, I'd be an idiot and would still be wasting most of my life at the keyboard. I found one can save days, weeks, months, or even years, by listening to those with experience. You don't seem to want to listen to the experience of many. Learn from other people's errors, you don't have time to make them all yourself ;-) Like others said, learn whats the right tool for the job, do the jobs you feel like and leave the others to the ones who like other jobs better. Don't forget to live. Like I like to say, I'm more scared of not living than of dying. I hope you can figure out the meaning ... I also find that messing about in languages I know nothing about tought me also about those I do know something about and also tought me what I don't want to do and what I'm wasting my time on with the wrong tool That's fine, but don't come crying to us in 3-4 years when you realize how much time you've wasted with such a capricious fetish. I'm still claiming that the world would have been a better place if computers hadn't been invented, or at least if user interface (cli, gui, whatever) has never been invented, and the more I learn the more I'm convinced on that matter ;-) I used to be more of a purist like you but after going through, c, c ++, java, matlab, perl, fortran (yes it's still alive and kicking ...), assembly, basic, pascal, logo, lisp and I don't know how many others I came to the concultion that if I can save three weeks programing, let the program run a couple of days instead of a couple of hours (assuming it needs to be run once or twice) and go out to date my wife, mountain bike, kite surf, watch the sunset or whatever, I'm much better of and sociaty is not the worst of it. If I do not change the reality, I will suffer till I'm dying; If I try to change the reality, maybe i will just suffer 20 years, and rest for my last 20 years. And the first place to change reality is YOUR PERCEPTION. There is no suffering here, other than your refusal to learn or use other languages. This is a task that, if you apply yourself, you will master in 6 months or so. The other task, that you propose, is not a 20 year task -- it won't be finished that quickly. So why suffer for 20 years when there's no reason to suffer for longer than a few months? As for resting the last 20 years, that assumes that there are enough people who value your work that they'll pay you enough that you will be able to rest, as opposed to finding whatever job you can to continue to support you. You've been given a plethora of reasons why programmers do NOT want a unified language. If they wanted one language, there are many programmers far wiser than anyone here that could have started that over the past few decades, but none of them have. Why? Because there is not only no need for one, but because it's counter productive. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 2:07 AM, Andrew M.A. Cater wrote: On Wed, Jun 24, 2009 at 01:43:35PM +0800, 明覺 wrote: What about XML, YAML, HTML, javascript, and such? No more browser? No more internet? :-) Of course I will use all of them, I even use windows vista everyday for playing games, that's my user role; for my programmer role, I will use Xml and html, for they are data files, not programming languages, I will try to write my own brower in C++, and use a subset of C++ to be the dhtml programming language instead of javascript. OK. I'll bite :) 15 years ago, I studied a couple of terms of evening classes in programming in C. One of the exercises thrown at us was: Build a reservation system for a 10 seat commuter aeroplane. Small, simple, defined - but harder than it looks on paper. Go for it: from your posts, you have the programming credentials. Try the following exercise: Build one in PHP / webforms (or Javascript) - web languages, anyway. Build one in pure Perl. Build one in C or C++ writing to a MySQL/Postgres database. Build one in C / C++ alone. Build one in assembler. Shouldn't take you long. That'll give you a much better feel for how/ why different approaches have their strengths and weaknesses. It will mean you're porting code that should be familiar and that you can read and understand because you wrote it. Its a limited problem: the real world is harder. For extra credit, put it up on a website somewhere and submit the URL back here for the code to be analysed by others here. And, in terms of usability and compatibility, since one is done for a browser, meaning a web front end, then be sure to write the others so they use a GUI. Sure, do them first from the command line if you want, but then when you're done, make them all work with a GUI. Come back when you're done. Then decide whether your ideas are feasible. I think this is a major point Andrew makes. This is a simple program, but, as I've pointed out, your comments indicate almost no experience in many parts of the programming world. Just trying to do this simple program in the ways he's suggested requires skills that the OP's comments make it clear he has not yet attained. Now if the OP had shown he's been there, done that, gone through all these kinds of experiences, I think many of us would take his comments MUCH more seriously. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 2:09 AM, 明覺 wrote: On Wed, Jun 24, 2009 at 1:21 AM, Micha Feiginmi...@post.tau.ac.il wrote: ... The only one language for microsoft is c#, oh wait, its visual basic, sorry wait a minute it's forms for the gui, assembly in the drivers in if you start digging you will find that half the system management is using scripts and batch scripts of one sort or another. I mean .net framework, I don't know they use scripts so heavily, but I think they are reducing the use of scripts, you even cannot find a command line console in vista. I can. Used it on my Mother's computer when running tests when Comcast, the local crappy cable/Internet provider screwed up. I know it's there because I've used it. What leads you to think it's not there? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 2:11 AM, 明覺 wrote: On Wed, Jun 24, 2009 at 12:36 AM, Micha Feiginmi...@post.tau.ac.il wrote: On Tue, 23 Jun 2009 09:18:16 +0800 明覺 shi.min...@gmail.com wrote: On Mon, Jun 22, 2009 at 10:18 PM, John Haslerjhas...@debian.org wrote: 明覺 writes: yes, currently it's true, but I hope one day I will be able to take full control of my system, and modify them as i like, if I have those other language programmed softwares installed in my system, it will be hard to maintain for me. If learning enough of another language to do maintainence is hard for you you aren't much of a programmer. Programming is not about knowing a language. Yes, language is just a tool, so I want to keep my tool simple and powerful, I do not want so many similar tools with the same functions. What you are saying is that you just don't want tools around. First of all they don't have the same function (and if you'd use them you'd know). And if you knew something you would know that c may be powerful but it is far from simple. There are things you can do in python in one line that you would need 100s of lines of code with c. 100s of lines of C code? how about drop the 100 lines into a function? 1) Not all libraries are bug free and producing such a library takes considerable effort and why does one need that effort when one can produce the same program quickly and easily in Python? 2) Perhaps it's not a specific function that will be re-used later, perhaps it's a combination of several things. 3) Considering that Python is written in C (or at least parts of it), if it worked better to include those functions in C, instead of in a new language, then the same code could have been done as a C library. How many people are trying to do that? 4) Do that. Then do it for EVERY Python function. And EVERY Perl function. And just try doing it with really high level languages. And see how HUGE a library you have -- it'll be so overburdened that nobody will be able to easily learn the API. People will stick with the original languages because it's easier to learn that language than such a complex library with an API. Now, here's one for you: You resist, in every possible way, learning any new languages and talk about how bad it is that you have to learn them. What languages do you know well? Which ones have you taken time to learn and to use for more than just a simple program? And what are your real objections to different languages and using them? The more I read your posts the more I get the impression that the issue here is you just want to justify not having to learn languages, which makes me wonder if you were hand-fed what it took to learn C and C++ and basically have never mastered any other language and don't want to -- and are going into more effort to justify that than what it would take to learn how to learn languages. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 8:05 AM, 明覺 wrote: On Wed, Jun 24, 2009 at 7:47 PM, Dale Harrisrod...@gmail.com wrote: On Wed, Jun 24, 2009 at 02:45:45PM +0800, 明覺 wrote: thank you! I thought they are kind to give me advice, but I'm wrong, they just want to laught at me, it doesn't matter, I finally know it, and maybe next time I will discover it earlier. I still insist my ideal: one language for computer programming! There is one language for all computer programming, it's the machine language. Everything reduces to that. It's all a bunch of ones and zeros. But then are you running on an Intel based platform, MIPS, ARM? Oh darn, different assembly languages. Face it, now speaking in metaphor, trying to insist on a hammer when you need a screwdriver will just make your life a living hell. You must be adaptable. you do not understand me, and I've been tired to explain my thoughts. thanks. Oh, we understand you quite well. Do you understand what we're saying? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 24, 2009, at 5:57 PM, Cowley Harris wrote: Hal Vaughan wrote. But I haven't seen anyone insult him yet -- but then I haven't read many of the overnight posts yet. Anybody here watch The daily show. The first time I ever saw a show, they had a piece where they had Donald Rumsfeld denying he had ever said there were weapons of mass destruction in Iraq, then they cut to old news footage of Rumsfeld saying We know they've got weapons of mass destruction, we know where they are. Classic. you're off your rocker Yes, a judgement call. Did you look at the context? Please, if you want to make me look bad, have the courtesy to quote me in context. The line was, All of US are telling you that you're off your rocker and on a fool's quest. It's a quick paraphrase of much of what has been said, and in context (which includes his question before it, which I'll leave to you to research), it's also idiomatic. I've never seen someone work so hard due to fear and sloth. That's not an insult. Look over his comments. I'll add that I've been corresponding with him privately as well. I'm not a counselor or LCSW or psychologist, but I have enough experience to recognize when someone is avoiding something from fear or trying to avoid work they don't want to do. I'll qualify myself if you need me to. You're not just ignorant Ignorant is a quantifiable judgement which fits because he has demonstrated an amazing lack of knowledge in the field in discussion. ...just plain ignorant and stupid Not an insult, a judgement call, but one that can be supported. Yes, maybe stupid can be an insult. IF context did not make a difference. I was calling the attitude just plain ignorant and stupid. That is not insulting the person. If you think so, then take teacher training and see why there is so much focus on addressing problems with a kid's actions, not with them. Now you can go on, but I stopped reading here because you pretty much took comments out of context to make your point. In other words, you tried to say that he makes me insecure and tried to focus on me insulting him when, what you've actually done, is exactly what you accuse me of. It's called projection: We like to say others are doing what we don't want to admit we're doing. The OP has made MANY statements that show he does not know his field. He's been given a LOT of good information, yet persists on saying that his way is right and that others don't understand him. That is, point blank, ignorant. I may not be smooth, I may not be graceful, but I'm not afraid to say, This is ignorant, when it is. And next time you want to accuse me or others of doing something, don't pull out just the parts that support your side and figure the context and parts that disprove it will be ignored or can be discarded. Feel free to misquote me more. I won't be reading it, though. I have a short tolerance for people that take my words out of context and try to say that they mean something other than what they mean in the full context in which they're used. In other words, at first I thought you might have had a point -- until you came on announcing, with the Jon Stewart reference, that you were going to seriously make me look like a full, but could only do so by removing the context of my statements. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 23, 2009, at 8:32 PM, 明覺 wrote: ... I open this thread as a programmer, you can ignore my questions about programming in the future, but you should not ignore my questions as a debian user. I don't know if your culture is aware of the story of The Boy Who Cried Wolf, but you might want to look it up and see what it says. The main point is that if people get used to seeing your emails following a pattern, after a while, they're not going to bother to read the same comments and lines of reasoning over and over if they have never found them interesting in the past. I think I'm talking about something interesting and serious, I hope there will be someone who is also interested in my thoughts, for those who ignore my questions, I can understand, not everyone will agree with me. You're talking about something serious, but, as many have pointed out to you here, and as I've pointed out to you several times in private emails, what you think is interesting is not really of interest to the rest of the programming world, and especially to those with experience -- and the more experience one has, from what I can tell, the less they seem to find that issue interesting. Most are not as much interested in your thoughts as they are in helping you see how you've boxed yourself into an area so small and esoteric that if you continue on your current path, your work will be of little interest to anyone but yourself. As to who is interested in your thoughts, it goes both ways. You continually reject any comments that disagree with you as valueless or as wrong and insist that you are right. If you treat others, especially those with much more experience than you, like that, then why are they going to be interested in your comments? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 23, 2009, at 8:34 PM, 明覺 wrote: On Tue, Jun 23, 2009 at 1:22 PM, Frank Lin PIATfp...@klabs.be wrote: On Mon, 2009-06-22 at 17:40 +0800, 明覺 wrote: On Mon, Jun 22, 2009 at 4:54 PM, Tzafrir Cohentzaf...@cohens.org.il wrote: On Mon, Jun 22, 2009 at 09:17:18AM +0800, 明覺 wrote: I want to keep the programs in my system all written in c/c++, no python or perl or any other programming languages, is it possible to reach it? I removed the 2 packages, python and perl, from my system, and of cause, I losed my desktop, is it possible to install a desktop manager without perl and python? which is the proper desktop manager? thanks Let's throw some data into this discussion: [..] Debconf needs perl. And you won't have much of a Debian system without debconf. I wouldn't try removing that one. I have quickly analyzed the files in my /usr/bin: 235 shell script 71 python programs 241 perl programs 986 ELF programs So does it mean debian determines the use of python and perl? could you help recommend some distributions that do not need perl or python? thanks Which company would hire a programmer that is unwilling to learn/use anything but C/C++? Your employability is low, and it will get lower and lower. Sorry for telling you the truth, but you can't live in a world with only C/C++ I'm looking for a job, and the title of my resume is Asm, C, C++ programmer, if I can find a job, I win; if I cannot, you win. :) Finding a job is a pain, but the real question is: Can you keep a job and not have to keep job swapping and continue to feed and clothe yourself and possibly provide for a family if you ever want one or continue, year after year, to pay your way in this world if you have trouble keeping the jobs you find. And as for looking for a job, do you want to share why you're currently out of a job? Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 23, 2009, at 9:13 PM, 明覺 wrote: ... I think the OP need to spend some time with Haskel, Erlang, Prolog, and Lisp. teaseOnly *then* can you truly learn to *hate* non C-based languages. ;)/tease Thank you, my plan is to first learn C/C++, then learn other languages, I learn other languages in order to extract out their advantages into C/C++, not for using them, of cause, before I have the ability to modify the g++ compilers as I like. The more I read in this thread, the more I wonder -- does anyone here have a link to a video of William Shatner delivering that famous three word line in a Saturday Night Live sketch from the '70s or '80s? I mean, seriously, GET A LIFE. It's just a programming language. GET A LIFE. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 23, 2009, at 9:51 PM, Dale Harris wrote: On Tue, Jun 23, 2009 at 09:28:10PM -0400, Napoleon wrote: You just don't get it, do you? There are different programming languages because there are NEEDS for different programming languages. There will NEVER be one programming language which meets all requirements. Just like there are hammers, screwdrivers, wrenches and so on to build things. Ah, but one is fabeled. Obviously you've never heard this ancient saying... One Language to rule them all, One Language to find them, One Language to bring them all and in the darkness bind them In the Land of Redmond where the Shadows lie. (just kidding folks. ;) No he's not. Hal (Who had the same thought in his mind on reading the idea of one language for all purposes.) -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 23, 2009, at 10:02 PM, 明覺 wrote: 2009/6/23 Jeff Soules sou...@gmail.com: I open this thread as a programmer, you can ignore my questions about programming in the future, but you should not ignore my questions as a debian user. Right now you are showing that you're a person who asks for advice, but does not listen to the response. People value their time and will not take the time to respond to someone like this, whether you're speaking as a programmer, a Debian user, an artist, or a fisherman. Don't waste people's time. Ever. yes i'm asking for advice, and I'm very happy to get so many good advices, and I'm trying to form a solution to include all the good advices, I'm not wasting other's time, we are just discussing and trying to figure out the best way. No, we're not figuring out the best way. All of US are telling you that you're off your rocker and on a fool's quest. You're saying, But you're wrong and I'm right. You talk about how different languages are just different ways to do the same thing. Well... okay... but you're writing to this list in English. From your sig and your name you're obviously a native Chinese speaker. Aren't English and Chinese just different ways to say the same thing? If you don't understand them both well, you might think so. But some things are much easier to do in one language versus the other. 飄飄何所似， 天地一沙鷗 -- in English, is it 'just the same thing?' It's not that Chinese is just better, there are plenty of things that are more natural in English than in Chinese. Just the same, if all you see in Perl is wrappers around C functions--if you think none of them bring new concepts [or clarity or simplicity] to C/C++ -- then you don't understand Perl. And you need to. Without lots of different ways of thinking about problems, you're like a frog in a well, saying look how small the sky is! A very good comparison -- human languages and programming languages. Then why we must have an official world language - English? What's the official language in the programming world? If you say you do not need an official programming language, then you are saying we do not need English to be the world official language, I believe no one will agree with you; if you say every programmer should learn many languages, then you are saying everyone should learn English, Chinese, French, oh, I believe everyone will hate you so much, I guess you are also a chinese, you should know how suffering we chinese have to learn English. Comparisons hold true on some levels, but few hold true on every level. In this case, you're taking one argument and stretching it beyond any boundary of logic or common sense. Yeah, I could go into it more, but why? You'll just say, But I'm right and you're not. I value every good concept in every language, No you don't. If you did, you'd understand the main message you've been told dozens of times. but please add that good concept to my familiar language, not force me to learn a new one; Nobody's forcing you to learn a darn thing. You don't have to do nothing -- except pay taxes and die (and I honestly don't know how taxes work in your country). You make it sound like a chore to learn a new language. For a true programmer it isn't. Learning a new language, for a real and true programmer, is and adventure. It's a chance to approach all problems from yet another perspective. I learned most languages in a few hours or days. When I first started looking at OOP, it took me a while, but once I got it, working with other OOP based languages was a snap. If you feel like you're being forced to learn languages, then you're in the wrong field. But after reading that line, I wonder Is all this because you have trouble with some languages -- it looks like you're essentially trying to go through all this so you don't have to learn languages you don't want to learn. I've never seen someone work so hard due to fear and sloth. or, I can reference another language so that I can improve my language, but please do not force me to use a new one. Nobody is forcing you to learn anything. You don't want to learn one, don't learn it. Quit the job -- but then when you want a new job, don't be surprised if they ask you why you quit that last one! Honestly, that you can even talk about being forced to learn a language, that you even have that as a concept in your brain, says even more about you. It tells us you don't want to learn something new. It tells us you don't want to explore. It tells us you see programming more as a chore than an art or challenge. It also says that we should have sympathy for whoever hires you as a programmer. The way computers working is simple, so there isn't any difficulties to implement a good concept in one language to another. Do you have any clue, when you make a statement like that, just how much it shows everyone that
Re: is it possible to install a desktop-manager without python and perl?
On Jun 22, 2009, at 8:00 AM, 明覺 wrote: ... Looks like a strange idea to me to run a one programming language only system, it would hint that there's a one fits all language and other are just for decoration purpose... (Well, some may agree I guess ;-) ) yes, currently, I'm almost a one programming language only people, I can accept the existence of other languages, but I think they should be optional, not necessory! Necessary for what?! For you to use a computer? It seems as though you're being unreasonable (on many fronts), but if you are not fan of certain software, then don't use it. Don't bitch about how it was developed. Those folks (eg., gnome, X.org, etc.) produced a product the way that they did and then offered it to the masses for free. cliche Beggers can't be choosers. /cliche you treat yourself a begger, I'm not, I'm a chooser. Happy begging to you! A rose by any other name is still a rose. You can call yourself a chooser, but your actions show you to be a begger. First, when you look at what's in even just a minimal Linux install, there is no way you're ever going to get through working on more than a few programs in the next few years. Second, when a programmer writes a program, if he has any wisdom (which is knowledge gained through experience, hence the more years, the more experience and the more wisdom), he will use the right tool for the right job. For instance, I need to use mainly Perl and Java, but have used many other languages. I find I can code 5x faster in Perl than Java and about the same, maybe better if I use Perl instead of C++. Hardly any of my Perl code is done as a wrapper for a C or C++ program. It is valid code that does a LOT of work and does it well. Since it's text processing, to do the same work in C or C++ would be a nightmare. If a program has to parse much user input or doesn't need extreme speed, or has any of a thousand other specs, then it is better to write it in Perl or Python and not C or C++. There are also a similar number of reasons to write code in C and C++. You lament there is no distro for someone with your need. That's because it's a self- imposed, self-limited need that no wise programmer or administrator will want to meet. They're too busy trying to get their work done so they can go see their girlfriend or get out and go ballroom dancing or meet a friend to go to the movies to stick with only one tool when there are a myriad of tools, each one working for specific jobs. If all you have is a hammer, then every problem is a nail, not a board to cut or a hole to drill or a nut to loosen. Limiting yourself to just a hammer when all those other situations exist is blindly unwise. You're doing the same thing by limiting yourself to only the C language family. You say you hate the other languages -- fine. If you want to cut off your nose to spite your face, do so. I hate Javascript, but I'm versant in it and use it when I need to. I sure as heck won't insist on wasting time writing an applet in Java, taking, perhaps, a few extra weeks, when I can code it in HTML and Javascript in much less time and have it work just as well. To do otherwise would be wasteful of my time and I have way too much to do with my life to spend it sitting in front of an LCD screen when I could be out with people and doing other things. You say you are a junior programmer, which means you are learning programming. Okay, LEARN. You've had advice from MANY people on this thread who are senior programmers or hackers or system admins. It's clear they all have significantly more experience than you and they're telling you, There is no point to what you're doing. Some are even pointing out how it will hurt your programming experience in the long run. It's your choice, but you've been well advised. There is a term for a novice who ignores the advice of those with experience and then proceeds to waste his time (the most precious commodity any of us have with a limited life span) trying to prove experience wrong and eventually learning the wiser ones know what they're talking about. That term is fool. I'm being blunt, but, honestly, I run a business on custom software I've written and I can do it because I learned from those who knew more than I did. If I refused to learn from people on this and other lists, I'd be an idiot and would still be wasting most of my life at the keyboard. I found one can save days, weeks, months, or even years, by listening to those with experience. You don't seem to want to listen to the experience of many. That's fine, but don't come crying to us in 3-4 years when you realize how much time you've wasted with such a capricious fetish. I hope you read all the posts here again with an open mind and learn from them. If not, well, then, it's your life you're throwing out and remember that in 5-10
Re: is it possible to install a desktop-manager without python and perl?
On Jun 22, 2009, at 9:18 PM, 明覺 wrote: On Mon, Jun 22, 2009 at 10:18 PM, John Haslerjhas...@debian.org wrote: 明覺 writes: yes, currently it's true, but I hope one day I will be able to take full control of my system, and modify them as i like, if I have those other language programmed softwares installed in my system, it will be hard to maintain for me. If learning enough of another language to do maintainence is hard for you you aren't much of a programmer. Programming is not about knowing a language. Yes, language is just a tool, so I want to keep my tool simple and powerful, I do not want so many similar tools with the same functions. Boy, I didn't realize that by junior programmer you meant you were that inexperienced in the field. I don't know if you realize that you've just basically said you are either unwilling or unable to understand the different reasons for different languages. EACH language is a tool, and each one is a DIFFERENT tool with a DIFFERENT purpose. It is rarely a whim why a programmer picks one language over another. There are often several, if not many reasons why one language is more appropriate and better for a job than another is. But there's no point in continuing any discussion. You've made it quite clear you're too busy being right to care what anyone more experienced has to say -- unless it's what you want to hear. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Jun 22, 2009, at 10:35 PM, 明覺 wrote: 2009/6/23 Napoleon rri0...@attglobal.net: 明覺 wrote: On Mon, Jun 22, 2009 at 10:18 PM, John Haslerjhas...@debian.org wrote: 明覺 writes: yes, currently it's true, but I hope one day I will be able to take full control of my system, and modify them as i like, if I have those other language programmed softwares installed in my system, it will be hard to maintain for me. If learning enough of another language to do maintainence is hard for you you aren't much of a programmer. Programming is not about knowing a language. Yes, language is just a tool, so I want to keep my tool simple and powerful, I do not want so many similar tools with the same functions. What you haven't learned is there are different languages FOR A REASON! No one language is best for everything. For instance - I can code web pages in C/C++ - but it is much faster for me to do it in PHP, Perl or Java. The same is true with anything else. I've got over 40 years of programming experience; in that time I've forgotten more languages than you have ever learned. Some no longer even exist. But every one of them had certain advantages and disadvantages - and those differences were a major reason why the languages were chosen for their particular projects. You don't like the way different languages handle strings - well, guess what. If they all did everything the same way, they wouldn't be different languages! To be blunt (like others) - so you don't like the fact different languages are being used on your system. There is no way you're going to be able to rewrite all that code in C/C++ in your lifetime. So you have two choices. You can accept that fact and continue to learn, using the tools available to you, no matter what language, just like the rest of us do. Or, you can continue to bitch about it and make yourself miserable. In this case, I suggest you try another profession - if you can't get over this little bit, you are not suited to be a programmer. This will just be the first of many frustrations for you. And one more thing - you can continue to bitch in this email list, but if you do, it won't be long before people will stop responding to you - for ANY post, even when you're asking for help. I open this thread as a programmer, you can ignore my questions about programming in the future, but you should not ignore my questions as a debian user. I don't know if your culture is aware of the story of The Boy Who Cried Wolf, but you might want to look it up and see what it says. The main point is that if people get used to seeing your emails following a pattern, after a while, they're not going to bother to read the same comments and lines of reasoning over and over if they have never found them interesting in the past. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Rogue Filename - Can't Do A Thing With It
On May 21, 2009, at 8:05 PM, Hilco Wijbenga wrote: 2009/5/20 Hal Vaughan h...@halblog.com: Recently I started getting errors from rsync on a machine I don't tend to have to log on to very often. I checked the bad directory and get this: [...@scarecrow:threshNet]$ ls -l reportX total 0 ?- ? ? ? ?? reportX/2009-r...@? snip/ [...@scarecrow:threshNet]$ rm reportX/* rm: cannot lstat `reportX/2009-raw\...@\037': No such file or directory I guess it will not work because rm doesn't work but you could try find . -type f -delete. Another command to try is unlink. No command worked with it as is. See the other reply of mine -- it took fsck to deal with it. Thanks for the ideas. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Rogue Filename - Can't Do A Thing With It
On May 22, 2009, at 12:52 AM, Kent West wrote: Hal Vaughan wrote: On May 21, 2009, at 8:05 PM, Hilco Wijbenga wrote: 2009/5/20 Hal Vaughan h...@halblog.com: Recently I started getting errors from rsync on a machine I don't tend to have to log on to very often. I checked the bad directory and get this: [...@scarecrow:threshNet]$ ls -l reportX total 0 ?- ? ? ? ?? reportX/2009-r...@? snip/ [...@scarecrow:threshNet]$ rm reportX/* rm: cannot lstat `reportX/2009-raw\...@\037': No such file or directory I guess it will not work because rm doesn't work but you could try find . -type f -delete. Another command to try is unlink. Thanks for the ideas. Tried both, here's the output for find: [...@scarecrow:ReportX]$ find . -type f -delete find: ./2009-raw?@: No such file or directory Got a similar message for unlink. Basically everything treats it as no file there. How about mc? Tried that, originally on ssh, from my iMac, but there was an issue because the iMac remaps the function keys. I know there's a way to turn that off, but I was going to have to re-attach a keyboard and screen to that computer anyway to run fsck, so I just waited to try it from a direct keyboard instead of remotely. MC didn't do anything the others didn't do. What did work was that fsck detected illegal characters in the filename, so the first ? (at least the first one) may have been unprintable. However, when fsck restored the filename, it had most of what I think was the original name, which was a lot longer. So my best guess is that the filename was corrupted and contained characters like backspaces in it. If this had been in my DOS 3.3 or ProDOS days, I'd have take out a sector editor and examined the file name that way and just altered it by hand. Sometimes I miss the simplicity of my old Apple //e. Thanks for the idea, though, The help is appreciated. Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Controlling Fonts and Font Sizes in GTK Applications
I have a Debian system hooked up to an HDTV. I'm using KDE as the DE and have had no problem adjusting the fonts used by KDE apps so they are big enough to see from across the room, but I'm stuck with itty- bitty fonts for Firefox and other GTK based apps. When I search with apt-cache, I find only gxset, which doesn't allow for setting fonts (at least not in Lenny). I've tried Google, but if it's obvious, I'm not using the right terms. What can I use to set the default fonts and font sizes for GTK programs? Thanks! Hal -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
