Re: forcing boot
On Jun 4, 2006, at 2:40 PM, Lawrence Horvath wrote: How can i force normal boot up even if the filesystem was uncleanly dismounted, i have a box that it looks like the HD is failing, but i still need some of the info off it, so i would like to get it to boot normally anyway so i can sftp/scp the files off it then i can replace the HD, but it refuses boot anything but single user mode. In single user mode /etc/netstart should bring up the network. If you have data on partitions other than the root, you can manually mount them read-only. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Linker Error: undefined reference to __ctype_b
On Mon, 10 Apr 2006, Premal Mishra wrote: Hi, Am tryin to compile a C program which uses 3rd party library. I get the following error during linking: undefined reference to __ctype_b and this error comes up for many more names. Whats may be the problem ? Without additional information, I'd say you are trying to link against a linux library. You are presumably running FreeBSD, so that is not likely to work. But, as that particular symbol can cause problems on linux if you have a library built against an older version of glibc, pehaps you are actually running Linux? If so, this is the wrong mailing list. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Perl 5.8.6 to 5.8.7 upgrade fails IPC tests
On Sun, 26 Jun 2005, Louis LeBlanc wrote: On 06/26/05 10:39 PM, [EMAIL PROTECTED] sat at the `puter and typed: On Sun, 26 Jun 2005 15:17:06 -0400 Louis LeBlanc [EMAIL PROTECTED] wrote: I'm having a little trouble with my perl upgrade from 5.8.6 to 5.8.7. -- cut -- ok 5 ok 6 semget: No space left on device you need to make disc-space, one idea is to run portsclean -C or to make some more space in /usr/home That would be the obvious cause, but not so: [...] Indeed. Disk space most likely has nothing to do with it. The semget(2) call returns ENOSPC when unable to allocate semaphores, due to hitting one of the many SYSV IPC limitations. Some things to look for: - You need to have SYSVSEM in your kernel, or sysvsem.ko loaded. - Check semaphore limits with ipcs -S - Check current semaphore usage with ipcs -s - Figure out what limit is being hit and do something about it :-) Some of the values can be tweaked at run-time, with sysctl kern.ipc.semxxx=value, others have to be set at boot time (in /boot/loader.conf), or compiled into the kernel. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Perl 5.8.6 to 5.8.7 upgrade fails IPC tests
On Sun, 26 Jun 2005, Louis LeBlanc wrote: On 06/26/05 02:32 PM, Mikko Tyljrvi sat at the `puter and typed: On Sun, 26 Jun 2005, Louis LeBlanc wrote: On 06/26/05 10:39 PM, [EMAIL PROTECTED] sat at the `puter and typed: On Sun, 26 Jun 2005 15:17:06 -0400 Louis LeBlanc [EMAIL PROTECTED] wrote: I'm having a little trouble with my perl upgrade from 5.8.6 to 5.8.7. -- cut -- ok 5 ok 6 semget: No space left on device you need to make disc-space, one idea is to run portsclean -C or to make some more space in /usr/home That would be the obvious cause, but not so: [...] Indeed. Disk space most likely has nothing to do with it. The semget(2) call returns ENOSPC when unable to allocate semaphores, due to hitting one of the many SYSV IPC limitations. Some things to look for: - You need to have SYSVSEM in your kernel, or sysvsem.ko loaded. - Check semaphore limits with ipcs -S - Check current semaphore usage with ipcs -s - Figure out what limit is being hit and do something about it :-) Some of the values can be tweaked at run-time, with sysctl kern.ipc.semxxx=value, others have to be set at boot time (in /boot/loader.conf), or compiled into the kernel. This certainly does make sense, but I'm not sure I'm actually running short here. I have SYSVSEM in my kernel (as well as SYSVSHM and SYSVMSG), and the relevant sysctls are: kern.ipc.semmap: 30 kern.ipc.semmni: 10 kern.ipc.semmns: 60 kern.ipc.semmnu: 30 kern.ipc.semmsl: 60 kern.ipc.semopm: 100 kern.ipc.semume: 10 kern.ipc.semusz: 92 kern.ipc.semvmx: 32767 kern.ipc.semaem: 16384 Ok, looks like the default settings, which are often too low for anything that makes heavy use of SYSV IPCs... root# ipcs -S seminfo: semmap: 30 (# of entries in semaphore map) semmni: 10 (# of semaphore identifiers) semmns: 60 (# of semaphores in system) semmnu: 30 (# of undo structures in system) semmsl: 60 (max # of semaphores per id) semopm:100 (max # of operations per semop call) semume: 10 (max # of undo entries per process) semusz: 92 (size in bytes of undo structure) semvmx: 32767 (semaphore maximum value) semaem: 16384 (adjust on exit max value) root# ipcs -s Semaphores: T ID KEYMODE OWNERGROUP s 655365432001 --rw---pgsqlpgsql s 655375432002 --rw---pgsqlpgsql s 655385432003 --rw---pgsqlpgsql ... such as databases :-) Have a look at /usr/ports/databases/postgresql80-server/pkg-message-server for some sample settings. Near as I can tell, this tells me I have at least 60 semaphores systemwide, 60 per id, 3 in use, none of which are being used by root (which is who I am running the test as). Shouldn't that leave 57 for the perl tests? Not necessarily. The SYSV IPCs is a particularly vicious piece of poor engineering. Semaphores come in sets containing one or more semaphore. With your settings you can have at most 10 sets, and a total of at most 60 semaphores, and at most 60 per set, and at most 30... something else. Also, at most 30 locks can be released in case a process unexpectedly exits. Easy, right? Looks like you'll have to use ipcs -sa to see the NSEMS column, which should tell you how many semaphores are in use. How many does it need to open? No idea. Read the code or just raise the retarded limits by a lot. Or try stopping postgres while running the tests. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: simple? sh problen
On Mon, 16 May 2005, Ian Smith wrote: Hopefully not too OT .. the only silly question being the unasked one .. How do I test whether a sh argument is an integer or not, so as to avoid failing on a syntax error from otherwise working code such as: [ $3 -lt 10 -o $3 -gt 600 ] echo $0 $1 $2: $3 invalid exit 1 when $3 is a non-integer argument? Do I need to delve into awk and REs, or is there something more simple I've missed in mans test, expr, etc? Here are some suggestions for functions to do the test: isnum() { expr $1 : '^[0-9][0-9]*$' /dev/null } isnum() { case $1 in *[^0-9]*|'') return 1;; esac return 0 } The second one is likely to be faster unless expr is a shell builtin (typically it it not). $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Encryption of login passwords--where and how is it done?
On Sat, 16 Apr 2005, Anthony Atkielski wrote: Where's the actual code that accepts the input of a password and/or encrypts it? I looked in login.c, but that only seems to call PAM or something; from that point on, I wasn't sure where to look. Start with crypt(3). I'm especially interested in knowing how a very long password (up to the FreeBSD limit of, I think, 128 characters) is hashed and mashed into an encrypted password, but I'm also generally interested in the whole process. I'd like to think that a 128-byte password consisting of random words and special characters would be just as secure as a shorter, completely random password, but that's only true if FreeBSD is hashing the entire 128-byte string in some cryptographically secure way in order to produce an encrypt password that is a function of every bit of the plaintext password. Look in /usr/src/lib/libcrypt/. The MD5 password hashing scheme is found in crypt-md5.c (the whole password is being used, btw). $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD 5.3 and too many files open...
Hi, On Wed, 23 Mar 2005 [EMAIL PROTECTED] wrote: Hi, i run into a problem on a FreeBSD 5.3-STABLE Applicationserver (2GB of RAM, 4GB Swap, Dual XEON 3.06Ghz). The box serves the xfrce4-panel for 80 Network Clients via ssh so the Users can start OpenOffice.org and firefox from that panel. It is goin very well, except a message (and also a problem) i run into the last day. If i tried to open firefox from a networkclient and received the message too many files open I saw that message on 3 different Network Clients and just after a fresh configure and the first time starting firfox on these Networkclients with that UID. I checked my configuration on the server about openfiles: kern.openfiles:9306 kern.maxfiles:65536 netstat -m 3952 mbufs in use 732/64000 mbuf clusters in use (current/max) [...] fstat shows me 393 on User A and on User B 3459. I did not check the other users. kern.maxfilesperproc: 5898 kern.maxusers: 384 My /boot/loader.conf looks like: kern.maxfiles=65536 kern.ipc.nmbclusters=64000 kern.ipc.nmbufs=256000 kern.maxproc=8192 kern.ipc.somaxconn=4096 So i do not get it, imho the configuration is just fine, but why do i get the message too many files open...? I think there is a separate limit for sockets, which you may be hitting. Check with sysctl kern.ipc | grep socket. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Xwrapper doesn't use .xinitrc, or does it?
On Mon, 21 Mar 2005, Andreas Davour wrote: I have just rebuilt my system and a few things aren't as they used to be. When I try to start X with Xwrapper no xterms or other windows appear. Isn't Xwrapper starting the stuff in .xinitrc like startx used to do for me (but wont do, since it no longer has the authority to do so)?? Isn't there some documentation for Xwrapper somewhere? You want startx(1). The .xinitrc file is read by xinit, btw. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Synaptics Touchpad driver
Hi, On Sat, 12 Mar 2005, Loren M. Lang wrote: It seems that FreeBSD 5.3 now has support in the kernel for the synaptics touchpad that my laptop has. Right now it's just running as a normal mouse, it looks like the support is disabled by default. In isa/psm.c, I can see the synaptics support in there, but it's disabled unless hw.psm.synaptics_support is set to 1. My question is how do I set it to one? It's setup as a TUNABLE_INT, but there is no sysctl for it. Does it only appear on boot? It is not a sysctl, it is a kernel tunable. You control it from the boot loader, for example by putting hw.psm.synaptics_support=1 into /boot/loader.conf. See loader.conf(5) and /boot/defaults/loader.conf for more information. $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Bus errors
On Sun, 6 Feb 2005, Tom Parquette wrote: Hi. I'm running 5.3-R with a custom kernel on a dual processor athlon. I've recently started getting Bus error showing up as a reason for various failures. e.g. setiathome, make buildworld. I've also gotten some panics (there are only about 5 messages that didn't tell me anything and (sorry) I did not write them down. I'm waiting for that to happen again. I've also seen some hangs where the screen saver (the blue rain one) hangs. For the hangs and panics I've only been able to get out of them with the hardware reset button. I've done some searching on bus error but it does not seem to apply. I only found entries talking about software interface problems. I'm starting to wonder if I have a hardware problem with this machine since it only seemed to start when I put a heavier load on it (e.g. setiathome) and the strange symptoms seem to be showing up in multiple places. The symptoms are typical of hardware problems. Could be bad memory, overly optimistic memory timings or CPU overheating. As you run setiathome, the latter would be my first guess. Try installing sysutils/xmbmon to keep an eye on the temperature of your CPUs. My Athlon 3200+ tends to fall over on its face when running setiathome on a hot summer's day, and there is no way I'm going to put even more loud fans in that box... $.02, /Mikko ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Where to find jw on FreeBSD?
On Sun, 5 Sep 2004, Matt Emmerton wrote: I've working on an open source project that recent went through a documentation frenzy and now we have a bunch of SGML (XML) docs that we reguarly convert to HTML and PDF. The person that usually does this runs RedHat and uses a project called 'jw' (jadewrapper) which is a nice front-end to all the docbook2xxx routines. I'm trying to do this on FreeBSD, and have docbook, jade and sgmltools installed from ports yet none of them include 'jw'. I know I can use the docbook2xxx routines, but was just wondering if there is a port that contains 'jw' or if this is a Linux/RedHat-only script? Sort-of. Google leads me to http://freshmeat.net/projects/docbook-utils/, which points to http://sources.redhat.com/docbook-tools/. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: [UPDATED]sick and tired of freebsd resolving problems
On Sun, 29 Aug 2004, horio shoichi wrote: On Thu, 26 Aug 2004 21:34:55 -0500 (CDT) Jorge Mario G. [EMAIL PROTECTED] wrote: -Original Message- From: Jorge Mario G. (snip) the problem is this: I CAN NOT RESOLV It's amazing this thread lives so long. So far, no new things are found, except for corrections of typos, beyond initial vague problem statement. What you are trying to do is damn simple thing, once it starts running. So, you are having dumb simple error somewhere. Now help us see the most elementary network characteristics of your machine. 1. Run the following script. Run it to the complesion, whatever happens. Just copy and paste your input and the machine responce. Don't try add anything, don't try remove anything, don't try change anything. % cat checknet.sh #!/bin/sh echo 0=$0 set -x ifconfig -a netstat -rn traceroute 216.136.204.21 traceroute freebsd.org echo done % 2. Run the following script for at least four seconds but not more than sixteen seconds. Just copy and paste your input and the machine responce. Don't try add anything, don't try remove anything, don't try change anything. % cat checkping.sh #!/bin/sh echo 0=$0 set -x ping 216.136.204.21 echo done % 3. Run the following script. Run it to the complesion, whatever happens. Just copy and paste your input and the machine responce. Don't try add anything, don't try remove anything, don't try change anything. % cat checkresolver.sh #!/bin/sh echo 0=$0 set -x nslookup freebsd.org nslookup freebsd.org 200.13.224.8 echo done % 4. Finally, if any change in /etc/hosts or /etc/resolv.conf, cat them. Just copy and paste your input and the machine responce. Don't try add anything, don't try remove anything, don't try change anything. If you still haven't found yourself what's wrong during those checkings, post the input and output. You forgot: 5. Run ipfw list (must be root) and 6. Run tcpdump -n udp port 53 at the same time as ping www.freebsd.org, for example in different windows (must be root to run tcpdump). $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: allowing users to mount cdrom
On Sat, 3 Jul 2004, jobse wrote: Dear List, When trying to mount the cdrom I get Operation not permitted. sysctl: vfs.usermount=0(what does that mean and how can I *permanently* change it to 1) I'd rather not set the sticky bit on mount/umount if I mustn't. suggestions? /jobse vfs.usermount allows non-root users to perform a mount, provided that they have sufficient access to both device being mounted and the mountpoint. Users won't be permitted to do other privileged operations, such as loading kernel modules, so in some cases the mount may still fail. To set vfs.usermount=1 on every boot, add it to /etc/sysctl.conf (see sysctl.conf(8)). One way to give access to assorted files and devices to the user currently logged in on the local console is to use /etc/fbtab (see fbtab(5)), thus: /dev/ttyv0 0600/dev/acd0 /dev/ttyv0 0755/cdrom When using some GUI based login doohickey (xdm, gdm, kdm, whatever), there are usually some kind of pre-session script where suitable code can be inserted to tweak permissions on things when someone is logging in on the local console (unless they are already reading /etc/fbtab... I dunno, I'm not using a GUI login thingy). $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: non-interactive password
On Sat, 24 Apr 2004, Louis LeBlanc wrote: Hey everyone. I'm trying to recreate a password script that sets passwords (as root on the local system only). The trick is that it must change the password non-interactively. I had this working, but when the system was trashed, the script was lost. Can someone refresh my memory on the command format? echo $password | pw usermod $user -h0 Might do the trick. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Perfoce client only
On Tue, 17 Feb 2004, Igor B. Bykhalo wrote: Is there a way to install the Perforce client only? We don't need the server, but i failed to find standalone client. And yes, command-line version is sufficient. There is no port but you can just download the command line tool from http://www.perforce.com/perforce/downloads/freebsd4axp.html. It is a single statically linked binary (p4) that does everything. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: mplayer problem
On Sun, 28 Dec 2003, Monah Baki wrote: Hi all, [ Lines wrapped at 72 chars ] I'm running the Freebsd 4.9 with mplayer 0.91 and mplayerplug-in for the ports tree and mozilla 1.4. I tried to view the spiderman trailer from apple's website, medium screem no problem, but when I try to choose large screen or full screen, I get a black display, and the following message (repeated 20 times) shows up on my terminal: ImageCodecbandDecompress cress=0xFFCE ( -0x32) -50 It works fine with medium screen. I've noticed the same thing. A cursory check with trafshow indicates that I simply don't have enough bandwith for the large version, which requires more than the 1.5Mbit my DSL connection will handle. The medium size feed usually stays below 1 Mbit. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Out of memory error
On Fri, 26 Dec 2003, Elliot Finley wrote: 4.8-R I have a process that needs to keep about 700MB of data in RAM. It crashes when it gets to about 512MB. I've looked for a sysctl variable to tune, but none of them jumped out at me. Same with man tuning' Any pointers would be appreciated. You'll need to tweak MAXDSIZ, see LINT. The default value seems to be picked up from /sys/OS/include/vmparam.h, and is 512MB on i386. There is a corresponding kernel tunable (at least in 4.9) called kern.maxdsiz you can put in /boot/loader.conf. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: spamassassin
On Sat, 20 Dec 2003, Doug Reynolds wrote: I've been trying to setup spamassassin on my freebsd box with postfix. I was reading the other thread about spamassassin. What i couldn't figure out, was how it got the mail from the mail system. I looked at spamd but it didn't look like the program. I googled and still couldn't find a good tutorial. If someone knows of a good tutorial and could give me a quick explaination, i'd be greatful. Oh, I am running 5.1-RELEASE. thanx Have you looked at ports/mail/spamass-milter? $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Ultra 2 SMP stop in probing devices during install
On Thu, 27 Nov 2003, Jens Baedeker wrote: Hi there, i'm pretty new with freebsd. Currently i run a couple of sparc20 with OpenBSD. But i want to give Freebsd 5.1 a try. Unfortunately i cant get this Ultra2 with 2 CPUs working. When boot with first install CDROM it recognises the 2 cpus during the bootprocess and start probing for devices. I can stop this probing but then i end up in the OBP prompt again. Checked google and the docs but nothing here. I updated the OBP to 3.11. You may have better luck with sparc specific questions over on [EMAIL PROTECTED] However, I strongly suspect that your Ultra 2 has a scsi controller that FreeBSD doesn't support (fas?), so even if you manage to boot, you'll have to run it diskless and configure it to netboot. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Something with port 53
On Sun, 9 Nov 2003, [ISO-8859-2] Gannater János wrote: Hello, I've posted this question to the freebsd-newbies list, but there nobody could answer it. Please help me with this qestion... I have a box with 2 IP addresses (with aliases). I usually get these error messages: Connection attempt to UDP IP_ADDR_1:1409 from IP_ADDR_2:53 What does this mean? It looks an awful lot like DNS replies from a server in IP_ADDR_2 to a client that has stopped waiting for a response and closed its socket. Pretty normal. I don't have any DNS server installed on the computer and the firewall is disabled on the system as well. Are you sure...? etc/hosts: ::1 localhost localhost.domain.com 127.0.0.1 localhost localhost.domain.com IP_ADDR_1 www.domain.com www IP_ADDR_2mail.domain.com mail etc/sysctl.conf net.inet.udp.blackhole=1 What about net.inet.udp.log_in_vain? $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Anyone using Linux-PAM on 5.x?
On Sat, 4 Oct 2003, Bill Campbell wrote: On Sat, Oct 04, 2003, Mikko Työläjärvi wrote: ... Having had some experience writing pam modules on a number of platforms, I whipped out my suite of pam test modules to have a look at this incredbible breakage you speak of. I mean, there is a specification for pam after all. Could you suggest where to look for information on writing modules to handle session tasks? I've looked at the PAM stuff superficially over the years, and would be very interested in doing something with the session modules to do things like automatically creating missing home directories with appropriate links to automounted directories. I haven't done much with this on FreeBSD yet since I need pam_ldap and nss_ldap support which only seems to be available in the 5.x tree. Well, it isn't exactly rocket science, just a small matter of programming. The pam interface itself is not very complicated; any decent programmer can pick it up in an afternoon. The specification can be downloaded from the OpenGroup, abd google will happily locate lots of sample code of varying quality for you. Modules aren't harder to port than any other code, the actual pam bits are mostly the same, modulo non-standard utility extensions, but those are usually trivial to port, replace or re-implement. One of the problems with pam is that there are very few guarantees on the context the module will be called in will look like. Will the uid be zero, or that of a user? Will all phases of the module be called from inside the same process? Will the session be closed twice after a fork? Will the service name stay the same over all phases? Trying to do anything more clever than verifying a password quickly leads to a jungle of special cases (number of pamified applications times number of platforms). Attempts at managing some kind of state between modules or different callbacks in the same module are more or less doomed to fail. Anyhow, just for kicks I grabbed the pam_mkhomedir module from RedHat9, and made it build on FreeBSD. It involved fixing the usual linuxisms (remove non-standard header files, add missing standard header files), and adding some FreeBSD-isms, like the leading dot. used for template files. A quick test with telnet shows that it seems to work, at least once :) The test also illustrates how the interaction between application and pam doesn't always work smoothly: the FreeBSD login program will already have discovered there there is no home directory, printed an error message and potentially denied access before the session modules are invoked. Thus there login messages contain: No home directory. Logging in with home = /. Creating directory '/home/testguy'. And then, as a band-aid, I threw in a chdir($HOME) in the session module to avoid having the user end up in /. Patch and build instructions below. Comes with no warranty. Slippery when wet. May cause bodily injury. Will probably break with NFS automounted home dirs. Etc... Enjoy, /Mikko Build: fetch ftp://ftp.redhat.com/pub/redhat/linux/9/en/os/i386/SRPMS/pam-0.75-48.src.rpm rpm2cpio pam-0.75-48.src.rpm pam-0.75-48.src.cpio cpio -id pam-0.75-48.src.cpio tar jxf Linux-PAM-0.75.tar.bz2 patch pam-0.75-mkhomedir-recurse.patch cd Linux-PAM-0.75/modules/pam_mkhomedir patch THIS_FILE cc -Wall -shared -fpic -o pam_mkhomedir.so pam_mkhomedir.c -lpam Patch: --- pam_mkhomedir.c.rh9 Sat Oct 4 22:10:49 2003 +++ pam_mkhomedir.c Sat Oct 4 22:14:51 2003 @@ -39,6 +39,8 @@ #include stdio.h #include string.h #include dirent.h +#include syslog.h +#include limits.h /* * here, we make a definition for the externally accessible function @@ -50,14 +52,15 @@ #define PAM_SM_SESSION #include security/pam_modules.h -#include security/_pam_macros.h +#include security/pam_appl.h +#define D(x) /* argument parsing */ #define MKHOMEDIR_DEBUG 020 /* keep quiet about things */ #define MKHOMEDIR_QUIET 040 /* keep quiet about things */ static unsigned int UMask = 0022; -static char SkelDir[BUFSIZ] = /etc/skel; +static char SkelDir[BUFSIZ] = /usr/share/skel; /* some syslogging */ static void _log_err(int err, const char *format, ...) @@ -151,15 +154,12 @@ mesg[0] = msg[0]; msg[0].msg_style = PAM_TEXT_INFO; - msg[0].msg = remark; + msg[0].msg = (char *)remark; retval = converse(pamh, ctrl, 1, mesg, resp); msg[0].msg = NULL; - if (resp) - { -_pam_drop_reply(resp, 1); - } + free(resp); } else { @@ -232,7 +232,8 @@ continue; /* We'll need the new file's name. */ - snprintf(newdest,sizeof(newdest),%s/%s,dest,Dir-d_name); + snprintf(newdest,sizeof(newdest),%s/%s,dest,Dir-d_name + + (strncmp(Dir-d_name, dot., 4) == 0 ? 3 : 0)); /* If it's a directory, recurse. */ if (S_ISDIR(St.st_mode)) @@ -351,7 +352,10 @@ if (stat(pwd-pw_dir,St) == 0
Re: Anyone using Linux-PAM on 5.x?
On Sun, 5 Oct 2003, Joe Lewis wrote: I know of the specification. However, I have a customized version of the pam-mysql module that I have tried to install. In a previous post to this list about a week or two ago, I wrote about getting only the following entrys in the log file (an example of one from today) : Oct 5 13:04:15 sharktooth login: in opempam_load_module(): no pam_mysql.so found Oct 5 13:04:15 sharktooth login: pam_start(): failed to load module Now, it is there in the /usr/lib/ directory, it has the correct permissions, correct owner, etc. In the /etc/pam.d/login file, I've even put in absolute paths, and non-aboslute paths. I've copied existing module entries, changing only the module name (such as pam_unix to pam_mysql). Looks like dlopen() fails. Is your module linked with any libraries other than libpam and libc? Such as libmysql... :) Unset LD_LIBRARY_PATH and check what ldd ./pam_mysql.so says (or read and grok the output of objdump -p). Anything listed as not found will result in exactly the error you are seeing, and the problem can be resolved using the normal linker tricks: symlink/copy or ldconfig or build with -R etc... Hmmm... odd; the modules all build and work just fine, modulo some minor tweaks mostly related to gcc 3.x, even though I have never built them on FreeBSD 5 before. Definitely no worse than when porting to certain commercial platforms. Can you at least point me to some documentation? I don't mind porting the module. I've copied the pam_unix.c module that comes with the openpam source code (dogwood release, the one in 5.x), changed references from pam_unix to pam_mysql, leaving the actual implementations in place so I make sure that I'm not messing something up, and STILL get the same log file entries, so it can't be the code in the module. What's the deal? I'm willing to bet on link problems. If I remember correctly one has to rebuild the openpam lib in order to enable debug logging, but it may be worth the effort if nothing else helps. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: explain annoying You have XXX mail messages
On Sat, 4 Oct 2003, Fernan Aguero wrote: +[ Mailing Lists Catcher [EMAIL PROTECTED] (03.Oct.2003 19:53): | | I don't know if this applies now but before I ran my own mail server I | used to see that sort of thing after I su'd so another account. For | instance if I logged in as user but su'd to a root it would tell me I | had mail but when I used the mail command it said no mail for user. I | forget how I was able to read root mail but I think it had something to | do with using su -m... Hi! Thanks for the reply. Yes I know this happens ... I can su to other users and see the 'You have mail.' message. However if I read mail for that user, and say, delete all messages, after I exit (/var/mail/user is now an empty file) and reenter I don't see this message anymore. | Anyhow root normally gets mail all the time from crontab and periodic | daily/weekly/monthly/security runs. | | Or Have a look in var/mail/root and others to see who has the mail. | Also use the mail command on the account you get the messages from next. I'm using the default sendmail that comes with freebsd. All mail to root gets redirected to me (alias root: [EMAIL PROTECTED]). And yes, I receive all the mails from root's cron jobs and periodic scripts. Still, /var/mail/fernan is empty all the time (I use procmail to deliver to ~/mail/inbox), and there's no /var/mail/root (since it is redirected to myseldf). All other users in the system have empty /var/mail/$USER files, since I'm the only user in the system :) So ... I'm still intrigued: i) I don't get 'You have mail.' but 'You have 50 mail messages.' That is to say ... the message is different, perhaps someone out there can identify the program that produces this kind of messages upon entering the shell? ii) ~/mail/inbox does not have 50 mail messages, it has thousands. Also, there are no new or unread messages ... iii) if I type 'mail', right after receiving the message, I get 'No mail for fernan.' as a reply. Thanks for any tip or suggestion, I'd guess that your .cshrc sets the mail variable to ~/mail or some other directory. This makes csh look for new mail in the specified path(s). See csh(1), under the description of the mail variable. All other programs use the environment variable MAIL, or default to /var/mail/$USER, where there is no mail for you. $.02, /Mikko Fernan | | | On Fri, 2003-10-03 at 17:08, Fernan Aguero wrote: | Hi! | | I'd like to understand why every time I open a new terminal | (or just type 'csh' in a terminal) I get a message similar | to: | | You have XXX mail messages. | | The default mailbox (/var/mail/fernan) is empty, since I | use procmail to deliver messages to several different | mailboxes under ~/mail. | | Still, I keep receving this message all the time. Right now | it says I have 50 messages. What are they? Where are they? | | I've already read the man pages for csh(1), looked in | sourced rc files (~/.cshrc, /etc/csh.cshrc, ~/.login, | /etc/csh.login) and, as far as I can see, there is nothing | that looks like checking on available messages. | | It's been also difficult to google a common phrase like 'you | have mail messages' to look for already answered questions. | | I apologize is this is a FAQ. | | Thanks in advance, | | Fernan | -- | Mailing Lists Catcher [EMAIL PROTECTED] | MGM Communications LLC kibserv | | +] -- F e r n a n A g u e r o http://genoma.unsam.edu.ar/~fernan ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: PASSWD file hosed...
On Sat, 4 Oct 2003, Eric F Crist wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello, I'm sure you're all getting sick of my being so vocal, but here goes anyway. I was messing around as root in KDE and found the KUser application. I tried adding a user (my girlfriend), and when done, nobody could log in. I managed Mmm... GUIs. to boot into single user, which logged me in a toor. I had to mount the file systems r/w and finally got my username's password reset, after finding a bunch of format errors in the /etc/master.passwd file. My problem is that the was no root user listed, just toor and all the others. How can I change If your machine has been up and running for a while, you will have backup files in /var/backup. Have a look at /var/backups/master.passwd.bak. If it looks contains everything you want, copy it to /etc/master.passwd. Otherwise just edit /etc/master.passwd and add the root entry. The default one without password looks like: root::0:0::0:0:Charlie :/root:/bin/csh Then run pwd_mkdb -p /etc/master.passwd and you're back in business. this? Also, I've set the console as insecure, so it's going to ask me for a root password. This is controlled in the file /etc/ttys. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Anyone using Linux-PAM on 5.x?
On Sat, 4 Oct 2003, Joe Lewis wrote: Question for you guru's; I've been trying to install a PAM module on my FreeBSD 5.1 system. Unfortunately, someone thought they were bright and included OpenPAM, which would be fine and dandy except for it is installed by default as static. This means I PAM is now AM, because nothing is pluggable. And the documentation on getting a 3rd party module to work is like slitting your wrists and doing pushups in salt water. Having had some experience writing pam modules on a number of platforms, I whipped out my suite of pam test modules to have a look at this incredbible breakage you speak of. I mean, there is a specification for pam after all. Hmmm... odd; the modules all build and work just fine, modulo some minor tweaks mostly related to gcc 3.x, even though I have never built them on FreeBSD 5 before. Definitely no worse than when porting to certain commercial platforms. My request is this... I'd like to know if someone has ported Linux-PAM (the same code from the old 4.x BSD) to the new 5.x OS. If so, do you have a copy of the ported code? If not, where do I post a copy of the stuff I am going to have to port so others can use it? Whatever difficulties you are experiencing, I'd say the reason is not specifically related to openpam or FreeBSD 5. Maybe if you post some details, such as source and error messages, someone will be able to help you. Regards, /Mikko Mikko [EMAIL PROTECTED] RSA Security ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: how to get system information
On Sun, 28 Sep 2003, Per olof Ljungmark wrote: I am wondering how to find out system hardware information from a running system, I know for instance pciconf(8) but what are the corresponding ones for memory, cpu etc? more /var/run/dmesg.boot to get the info at boot time. vmstat 5 5 will give you 5 items 5 sconds apart to show you procs, memory, page, disks, faults and cpu info. Thanks, that was a good start. Now, if I wanted to see more detailed info on the processor than the dmesg.boot output: Origin = GenuineIntel Id = 0x673 Stepping = 3 Like cache size etc.? You should be able to get some cpu info with ports/sysutils/x86info. Dunno about memory (its physical configuration, that is). USB devices can be listed with usbdevs, pci ones with pciconf, ATA devices with atacontrol, SCSI ones with camcontrol, XFree prints lots of info on your graphics card into /var/run/XFree86.0.log. Also, some of the values detected at boot end up as sysctls (usually under hw). I know of no simple way to get complete hardware information on a silver platter, though. Nor do I really care :-) $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Problems running new diablo-{jre,jdk} JVMs
On Tue, 2 Sep 2003, Kirk Strauser wrote: I've installed the new diablo-jdk and diablo-jre ports on my recently-upgraded-from-4.x FreeBSD 5.1 server. Whenever I try to run any of the JVMs, either by calling them directly or via the javavm script, I get this error: [EMAIL PROTECTED]:/lib# /usr/local/diablo-jdk1.3.1/bin/java -version /usr/libexec/ld-elf.so.1: /lib/libm.so.2: Undefined symbol __fpclassifyf I'm at a loss. Any ideas? It won't work on 5.1. See: http://www.freebsdfoundation.org/downloads/java-errata.shtml $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: zmore for bzip2?
On Sun, 31 Aug 2003 [EMAIL PROTECTED] wrote: Adapting zmore for the case where you specify the files to display on the command line is not problem at all: diff kk zmore 5,14d4 get_decompressor () { case `file ${1--} | sed s/[^:]*: *\([^ ]*\).*/\1/` in compress*) DECOMPRESSOR=uncompress -c;; gzip*) DECOMPRESSOR=gzip -cdfq;; bzip2*)DECOMPRESSOR=bunzip2 -cdq;; *) DECOMPRESSOR=cat;; esac } 56,57c46 get_decompressor ${FILE} ${DECOMPRESSOR} $FILE | eval ${PAGER-more} --- gzip -cdfq $FILE | eval ${PAGER-more} But when zmore is used as a pipe or with input redirection, things become more complicated. gzip -cdfq | eval ${PAGER-more} In order to detect the type of data passed on STDIN, the get_decompressor function or any other means of detection would consume STDIN. STDIN, however must be passed to the decompressor after the type of data has been detected. I don't have an idea hot to 'duplicate' STDIN, so it could be As a somewhat ugly hack, you could read a few bytes, match against known magic numbers and then prepend those bytes to the stream before feeding to the decompressor, for the cost of an extra cat process: magic=$(dd bs=1 count=3 2/dev/null) case $magic in BZh) DECOMPRESSOR=bunzip2;; *) DECOMPRESSOR=gzip -cdfq;; esac (echo -n $magic; exec cat) | $DECOMPRESSOR | eval ${PAGER-more} Note that gzip will handle compressed data and gzipped data as well as plain text. $.02, /Mikko consumed twice. Sure, writing it to a temporary file would be a workaround: 38,53c28 FILE=/tmp/.zmore.${$} touch ${FILE} 2/dev/null if [ ${?} -ne 0 ]; then echo can't create temporary file exit 1 fi chmod 0600 ${FILE} cat ${FILE} if [ ${?} -ne 0 ]; then echo can't create temporary file rm -f ${FILE} exit 1 fi get_decompressor ${FILE} cat ${FILE} | ${DECOMPRESSOR} | eval ${PAGER-more} rm -f ${FILE} --- gzip -cdfq | eval ${PAGER-more} But I don't like that. I myself sometimes work with compressed files larger than anything I would be happy to write to /tmp or somewhere else (even though those cases usually rather use zcat and its cousinds than zmore...) Kurt On Sat, Aug 30, 2003 at 12:21:45PM -0400, Chuck Swiger wrote: David Kelly wrote: [ ... ] Yes, of course. But zmore is smart enough to figure out what to do with several compression techniques, or even to handle non-compressed files very trivially and without hassle. 'zmore' is a simple shell script which calls gzcat | ${PAGER-more}. One solution to your problem, or at least a solution, would be to change zmore to look for a trailing bz/bz2 or invoke bzcat instead. Another would be to change the sources of gzip to recognize the bzip2 magic files bytes, extending the detection of gzip versus classic LZH used by compress. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Off Topic Regex Question
On Thu, 28 Aug 2003, Roger Williams wrote: I know thins is not the place but I know one of you know this one off the top of your head. I have: $list = dog 1 1 1 cat 2 1 snake 111 and I want to end up with: dog 1 cat 2 snake 1 I thought $list =~ s/ \d \d/ \d/g; would do the trick, but that gives me: dog d 1 1 cat d snake d 1 I'm not sure I understand your criteria, but try something like: $list =~ s/(\w+\s+\d\s?)[\d\s]*/$1/g; $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dumping/reading memory space of a running program.
On Mon, 25 Aug 2003, Sten Daniel Sørsdal wrote: How do i read/dump the entire memory space of a running program? Is this possible? I really dont have the option of restarting it. gcore(1) perhaps? $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: rpc.statd and NFS Server on 4.8-RELEASE
On Sat, 19 Jul 2003, Ernest H. Rice, III wrote: Folks: Excuse the interruption, but I have a probably very stupid question... I enabled NFS Server facilities on my 4.8.2 FreeBSD system recently. I export ONE directory, and when nfs is started (via the normal booting process) my internet connectivity slows down. Additionally, my system can run out of swap space (I have 128 MB RAM and 512 MB swap). I never ran out of swap before on FreeBSD and this troubled me. I ran top and saw rpc.statd usurping about 257M of memory. When I disable NFS Server facilities, and reboot, my internet performance returns (and - of course - rpc.statd is no longer running. Also, I no longer run short of swap. When I export the directory I have one line in /etc/exports which just lists the directory to be exported. What am I doing wrong to have rpc.statd using SO MUCH memory. It isn't. See the FAQ. http://www.freebsd.org/doc/en_US.ISO8859-1/books/faq/admin.html#STATD-MEM-LEAK $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: looking for http/web copy program - I think
On Sat, 19 Jul 2003, Robert Chalmers wrote: Is there a FreeBSD based app that I can use to bulk copy about 800 .rm files from one site to another? I really don't want to have to do it one at a time. ftp isn't available. The only access I have is via the web port. bummmer but theres no choice You could, for example, use wget (ports/ftp/wget). Never mind the ftp category - it speaks http too. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: how to copy just part of a file?
On Wed, 16 Jul 2003, BSD baby wrote: Is there an easy built-in way to copy only part of a file? I want to take a WAV audio file and copy from #__ bytes to #___ bytes into a new file. (I'm making 30-second clips of files.) Try dd. Thus: dd bs=1 skip=$offset count=$length infile outfile It is possible that a block size of one byte is kind of inefficient, so if you can operate in units of some larger size that might be good. For example bs=1k. See man dd. Also, remember that WAV files have a 44 byte (or was that 42...?) wav header. Yes, you can extract it with dd bs=44 count=1, but I'm too lazy to find out whether you can just prepend it to any split files right now (i.e. if the file contains any length info; I suspect not). Though I found a scripting way to do it with PHP, I'm wondering if there's a more direct way to do it with basic GNU/BSD commands. Ever heard of POSIX... ;-) $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Hi FreeBSD forum's Champs
On Sun, 13 Jul 2003 [EMAIL PROTECTED] wrote: Hi Forum I had a task to capture Windows dump (screen shot) in different formats. xwd -window image.xwd I also want to convert this default format (xwd) to X window bitmap, X window Pixmap. Any suggestion / utility to achieve this If you install the netpbm package or port, you can do things like xwdtopnm image.xwd | ppmtoxpm image.xpm and xwdtopnm image.xwd | ppmtopgm | pgmtopbm | pbmtoxbm image.xbm There are lots of other possibilities as well; browse through /usr/ports/graphics or http://www.freebsd.org/ports/graphics and you'll see... $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: mod_ssl question: using my own CA?
On Sat, 12 Jul 2003, admin wrote: OS: FreeBSD 4.8 apache 1.3.27 modssl 2.8.14 goals: generate a server.crt file for apache generate a server.key file for apache I will be my own CA Hi, okay I am trying to find a way to overcome this most elusive and vague documentationt that I am finding on the modssl.org website. I am completely confused by the documentation at this point. from: http://www.modssl.org/docs/2.7/ssl_faq.html#ToC29 So a script named sign.sh is distributed with the mod_ssl distribution ^ --- snip 4. Now you can use this CA to sign server CSR's in order to create real SSL Certificates for use inside an Apache webserver (assuming you already have a server.csr at hand): $ ./sign.sh server.csr This signs the server CSR and results in a server.crt file. shell# find / -name sign.sh % tar ztf mod_ssl-2.8.14-1.3.27.tar.gz | grep sign.sh mod_ssl-2.8.14-1.3.27/pkg.contrib/sign.sh $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: extracting cd-audio to wav
On Fri, 11 Jul 2003, Benjamin Walkenhorst wrote: Hello everybody, Is there some kind of program to digitally read audio-cds to wav-files under FreeBSD? Yes. Under GNU/Linux I use cdparanoia, which I have I've come to like very much, but cdparanoia III release 9.8 (March 23, 2001) refused to compile, using both FreeBSD's make and gmake using FreeBSD 5.0-RELEASE (I later upgraded to 5.1 by installing the new source-tree and recompiling the system by make buildworld + make kernel KERNCONF=MYKERNELCONFIG + reboot + make installworld, but the problem remains). I looked up the FAQ and found out cdparanoia doesn't run on FreeBSD currently. Can I use Linux binary emulation to run cdparanoia from my linux installation (/ and /usr are ext3, so I can read them from FreeBSD)? Or is there any program for FreeBSD which offers error correction similarly to cdparanoia? Or any cd-ripper at all? Don't believe everything you are told. cd /usr/ports/audio/cdparanioa; make install clean It is version 3.9.8 and it works just fine on somewhat recently cvsupped system. If you didn't use the port, I can understand if it did not build - the port has some 1k lines of patches. There is also dagrab and cdda2wav (the latter is in the sysutils/cdrtools port). Additionally, the ata driver has track nodes that you can dd from :) $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: IglooFTP in free(): error: chunk is already free
On Sat, 6 Jul 2003, Edy Lie wrote: anyone install iglooftp from ports and run it on KDE3.1 ? For some reasons when i tried to launch it from command line, i am getting that error message. Any idea ? It means there's a bug in IglooFTP. Put the following patch in /usr/ports/ftp/IglooFTP/files/patch-ad and rebuild the port, then things should work better. $.02, /Mikko patch-ad: --- dir_tree.c.org Sat Jul 5 14:26:53 2003 +++ dir_tree.c Sat Jul 5 14:27:02 2003 @@ -153,7 +153,6 @@ new_ptr[0] = 'B'; gtk_ctree_node_set_row_data_full (ctree, parent, strdup (new_ptr), (GtkDestroyNotify) free); free (new_ptr); - free (ptr); // verify that this is a good thing to do PARENT_BROWSED_FLAG = TRUE; } ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: IglooFTP in free(): error: chunk is already free
On Sat, 6 Jul 2003, Edy Lie wrote: After creating the file, the moment i start make install it gives me the following error message su-2.05b# make install === Patching for IglooFTP-0.6.1 === Applying FreeBSD patches for IglooFTP-0.6.1 Ignoring previously applied (or reversed) patch. 1 out of 1 hunks ignored--saving rejects to FTP.c.rej Patch patch-aa failed to apply cleanly. *** Error code 1 Stop in /usr/ports/ftp/IglooFTP. Do you have any idea ? make clean; maybe cvsup; try again. It worked a few hours ago. PS. if it is a patch ... should we notify the port maintaner? Thank you. Sure. I don't use IglooFTP, but once you manage to figure out if the one-line patch solves your problem, feel free to file a PR and watch it rot. $.02, /Mikko On Sun, 2003-07-06 at 05:33, Mikko Työläjärvi wrote: On Sat, 6 Jul 2003, Edy Lie wrote: anyone install iglooftp from ports and run it on KDE3.1 ? For some reasons when i tried to launch it from command line, i am getting that error message. Any idea ? It means there's a bug in IglooFTP. Put the following patch in /usr/ports/ftp/IglooFTP/files/patch-ad and rebuild the port, then things should work better. $.02, /Mikko patch-ad: --- dir_tree.c.org Sat Jul 5 14:26:53 2003 +++ dir_tree.c Sat Jul 5 14:27:02 2003 @@ -153,7 +153,6 @@ new_ptr[0] = 'B'; gtk_ctree_node_set_row_data_full (ctree, parent, strdup (new_ptr), (GtkDestroyNotify) free); free (new_ptr); - free (ptr); // verify that this is a good thing to do PARENT_BROWSED_FLAG = TRUE; } -- In the windoze world, I am limited by the tools that I can use, In Unix, I am limited by my own wisdom. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: using bind() call on FreeBSD
On Mon, 23 Jun 2003, Gagan Grewal wrote: Hi Folks :) I am trying to write a simple a server process which follows this sequence... socket() Assuming a struct sockaddr_in addr; around here somewhere, do: memset(addr, 0, sizeof(addr)); addr.sin_port = htons(blah); etc ... I.e. zero-fill the whole struct before use. $.02, /Mikko bind() listen() accept() . . . close( descriptor from accept() ) close( descriptor from socket() ) But I am getting error 99 (Cannot assign requested address) from bind(). I am trying to bind the socket on 127.0.0.1: (This works on Linux though) Are there any special/extra things I need to do in /etc or elsewhere to make this program run on FreeBSD ? I am running FreeBSD 4.7-RELEASE Any help/pointers from you folks would be great :) Thanks in advance :) Regards, Gagan. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: accidently deleted user www
On Mon, 23 Jun 2003, Andre wrote: Hi i accidently deleted the user www and now my apache doesnt start. is there a fast way to solve this problem? Don't do that. :-) If your machine has been running for a while there should be backups in /var/backup, so you can do: # grep '^www:' /var/backups/master.passwd.bak /etc/master.passwd # pwd_mkdb -p /etc/master.passwd $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: flashpluginwrapper for Flash 6 (Shared Library Tools redux)
On Thu, 19 Jun 2003, Joe Kelsey wrote: Ok. I hand-modified the linux-flashplugin6 shared library to remove all of the DT_NEEDED entries. I then modified flashpluginwrapper to add the following functions: [ ... snip wrapper functions ... ] I installed this new version of flashpluginwrapper, installed the linux-flashplugin6 library in browser_plugins. Now, the Flash 6 code segfaults in pthread_mutex_init(), called from Program received signal SIGSEGV, Segmentation fault. 0x284549d6 in pthread_mutex_init () from /usr/lib/libc_r.so.4 (gdb) bt #0 0x284549d6 in pthread_mutex_init () from /usr/lib/libc_r.so.4 #1 0x298016b0 in MPCriticalSection::MPCriticalSection () from /usr/X11R6/lib/browser_plugins/libflashplayer.so [ ... snip rest of stack trace ... ] Anyone familiar enough with Mozilla internals to understand why pthread_mutex_init segfaults? Has Mozilla started its own threads by the time it loads the libraries? I don't know where to look to track this down. Maybe a difference between Linux and FreeBSD in the pthread semantics? If the flash plugin is making direct calls to pthread functions, you'll probably have to provide some sort of compatibility wrappers for those as well, as the pthreads interface make use of lots of implementation specific data types supplied by the caller. These types are not likely to have the same definitions in Linux and FreeBSD. Check the documentation of all pthread* functions referenced in the plugin, then find out the definitions of the types of their arguments from pthread.h on each platform. For example: pthread_mutex_init(pthread_mutex_t *mutex, const pthread_mutexattr_t *attr); On FreeBSD, pthread_mutex_t and pthread_mutexattr_t are pointers, on Linux they are structs. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Problem with USB ulpt0 and CUPS
On Wed, 18 Jun 2003, Thomas T. Veldhouse wrote: I posted this in April and received no response. However, this has been an ongoing issue since at least 2001 (where I found the first reference to this trouble via Google). The problem seems to be that the FreeBSD USB LPT driver (even with no-reset) is somehow dropping the first bits of the data stream, causing a page full of trash to be printed prior to the actual print job. I have no verified this problem with 5.1-RELEASE as well. FWIW, I had similar problems with USB printer and 4.7. Inserting a sleep(1) between opening the device and the first write proved to be a sufficient workaround. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: xterm questions
On Sun, 15 Jun 2003, Alex de Kruijff wrote: I have a configuration question about how to change the default selecting proces. I would like it when i select a URL (with a double click), that it select the whole URL instead of just one word (sperating at [:/@]). It this posible? Try running: xterm -cc 46-47:48,58:48,64:48 or put something like: XTerm*charClass: 46-47:48,58:48,64:48 in your Xdefaults. See xterm(1), especially the section on character classes. Since URLs can contain all sorts of junk, you may want to expand on the above example. $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Problems with my terminal device files
On Sun, 1 Jun 2003, Daniela wrote: [...] I can't allocate new pty's, no matter what I do. [...] Xterm says that I don't have enough ptys. How do I make more? Umm... this is on 4.x, right? Have you tried making more pty device nodes? Thus: # cd /dev # sh MAKEDEV pty1 # 32 more nodes... # sh MAKEDEV pty2 # another 32 nodes... # sh MAKEDEV pty3 # etc... $.02, /Mikko ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: /usr/src on a dedicated drive?
On Sun, 16 Mar 2003, Brian McCann wrote: Hey all. I've got a problem on one of my systems (lack of drive space) that forced me to remove /usr/src. However, I like that system to track the STABLE version. I don't think it'd be a problem, but I'd like someone else to confirm this. Would there be any problems if I put /usr/src on it's own disk? No. :-) /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: DHCP Server learning name servers since server itself is dhcp'd??
On Sat, 8 Mar 2003, IAccounts wrote: Alternatively is there a way to dynamically tell BIND to get it's forwarders list from /etc/resolv.conf? Here is a shell script snippet that I use on my laptop. It gets called from make_resolv_conf() in /etc/dhclient-enter-hooks, where I make sure not to overwrite /etc/resolv.conf (it always points to localhost). Extracting nameserver addresses from resolv.conf is trivial, though. A prerequisite is that the forwarders clause in named.conf is on a single line by itself, for example: forwarders { 192.168.250.254; }; 8 LOGGER=echo named_conf=/etc/namedb/named.conf # Args: one or more nameserver IPs update_forwarders() { address_list= for nameserver in $* ; do test X$nameserver = X127.0.0.1 continue # Stupid server... address_list=$address_list $nameserver; done address_list={ $address_list }; sed_command='/^options/,/^}/s/\([^#\/]*\)forwarders.*/\1forwarders' sed_command=$sed_command $address_list/ sed $sed_command $named_conf $named_conf.dhcp if cmp -s $named_conf $named_conf.dhcp; then : else $LOGGER New DNS servers: $* if [ ! -f $named_conf.org ]; then cp $named_conf $named_conf.org fi cp $named_conf.dhcp $named_conf ndc reload fi return 0 } 8 It is more elegant in perl, but dhclient-enter-hooks is a shellscript, so it felt easier to just add it there. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: SBC DSL vs Fbsd 4.7 PPPoE?
On Mon, 3 Mar 2003 [EMAIL PROTECTED] wrote: How-do. Has anyone gotten SBC's DSL service working on FreeBSD? Yes. Well, it was called PacBell DSL at the time, but it is still working. I pretty much just used the pppoe entry from /usr/share/examples/ppp/ppp.conf.sample. Something like this: /etc/ppp/ppp.conf: sbc: # Replace rl0 with your interface connected to the DSL box set device PPPoE:rl0 # Hmmm... maybe these aren't needed anymore: set mru 1492 set mtu 1492 set speed sync enable lqr set cd 5 set dial set login set redial 0 0 set authname set authkey add default HISADDR /etc/rc.conf: network_interfaces=rl0 lo0 # The pppoe interface has to be configured as up. Nothing else. ifconfig_rl0=up ppp_enable=YES ppp_mode=ddial ppp_nat=YES ppp_profile=sbc I've tinkered with netgraph, pppoe, and my ppp.conf file but still get an error about device tun0 or tun1; I'm not sure what to do. Any chance that you might be a bit more specific? (As in: exactly what you did, and exactly what error messages that resulted in. FreeBSD version might be useful information too). Both tun and netgraph should be autoloaded by ppp(8). Everything should work out-of-the-box. No tinkering required :-) The only gotcha I can remember was that the interface used for PPPoE has to be ifconfig up in order to be usable. I don't even know whether that is true any more. Also, how do I know what hostname to set in my rc.conf file ... before I've been assigned a dynamic IP?? Anything you want. SBC won't care and it won't be visible to the outside world anyway. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: Xfree86-4 question re OpenGL
On Mon, 24 Feb 2003, Murray Taylor wrote: I installed XFree86 4.2.0 from a current ports tree WITHOUT having ever had XFree86-3 loaded. [...] When I run it from command line i get this ttyp0 kpovmodeler Xlib: extension GLX missing on display :0.0. Xlib: extension GLX missing on display :0.0. PMPart: TreeView added PMPart: DialogView added PMPart: GLView added PMPart: GLView added PMPart: GLView added PMPart: GLView added and the GUI has the 4 display windows showing No OpenGL support I have Mesa 3,4,2_2 loaded ( it appearesd during the KDE build I think ) How cam I get the OpenGL support working? From my /etc/X11/XF86Config: Section Module ... Load GLcore Load glx ... EndSection That may or may not be it. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: catch output from sh -xv to a file
On Sun, 16 Feb 2003, Per olof Ljungmark wrote: Need advice on how to modify first line of shell script so that script can run normally (as called from another program, not from the command line) and write debugging output to a file like: Short answer: you can't. Not if you absolutely want to change just the first line, and not really if you want the script to run normally, as you would have to divert stderr. #!/bin/sh echo DEFANGED.1007 exit #!/bin/sh -xv pipe stuff to file Try: #!/bin/sh exec 2 /tmp/logfile; set -xv # rest of script here... Should work, except for the script being unable to print stuff to stderr. There is also the possibility that I'm totally missing your point... :) $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: #!/bin/sh execve
On Sat, 8 Feb 2003 [EMAIL PROTECTED] wrote: say i have 2 scripts, scriptA and scriptB. scriptA --- #!/bin/sh ./scriptB 1 2 3 scriptB --- #!/bin/sh echo 0:$0 echo 1:$1 echo 2:$2 echo 3:$3 -- $ ./scriptA $0:./scriptB $1:1 $2:2 $3:3 -- according to execve(2), only a single [arg] should be recognized: #! interpreter [arg] When an interpreter file is execve'd, the system actually execve's the specified interpreter. If the optional arg is specified, it becomes the first argument to the interpreter, and the name of the originally execve'd file becomes the second argument; otherwise, the name of the originally execve'd file becomes the first argument. The original argu- ments are shifted over to become the subsequent arguments. The zeroth argument is set to the specified interpreter. so the argv[] array in execve() should be loaded as: argv[0]=sh, argv[1]=scriptB, argv[2]=scriptA, and argv[3...]=command line args passed to scriptA. i read many many execve() man pages, and it seems like this is the way things should be. but in practice, it appears on many unix's, argv[] gets loaded additionally with any options given to a script (which is given as the [arg] to the interpreter) on the 1st line of a script. can anyone tell me if this is proper, and why or why not? there doesn't seem to be consistency across unix's. some ignore, or give an error if more than one [arg] exists on the 1st line of a script. The only thing I can find in IEEE Std 1003.1-2001 (aka SUSv3) is If the first line of a file of shell commands starts with the characters #!, the results are unspecified. which would indicate that there is no proper way of doing this. You may also want to have a look at bin/16393; at the bottom is a list of how some unices handle the situation. Your best bet at trying to be portable is to use at most one argument, no whitespace and no #. The PR: http://www.freebsd.org/cgi/query-pr.cgi?pr=16393 $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: LINUX sysinfo syscall
On Sun, 12 Jan 2003, William Gianopoulos wrote: This has probably been asked before, but I could not find any info searching the archives. I am trying to run the Linux version Tapeware from Yosemite under FreeBSD 4.1. It fails because the Linux syscall sysinfo is not implemented. My questions are: 1- Is there some other port/package or option I should be using? 2- Would a later version of FreeBSD fix this? Linux sysinfo was checked into the source in mid-2001. Upgrading should help. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: linux compatability
On Sat, 28 Dec 2002, Adam K Kirchhoff wrote: Hey folks, I'm noticing some odd behaviour with the linux compatability recently. I have this small gnome app called gnome-run. It links against a number of gnome libraries that I've copied from my linux partition over to /compat/linux and put in the appropriate directories. The problem is, when executing this app, the linker seems to try and load the libaudiofile.so.0 file from my FreeBSD installation instead of my /compat/linux installation. When running gnome-run it only does this for that one library (though other apps have shown this to be a problem with other libraries such as libgmodule-2.0.so.0, libglib-2.0.so.0, and libgobject-2.0.so.0) This is what happens when I try to launch gnome-run from an xterm: [ adamk@sorrow ~ ]$ gnome-run gnome-run: error while loading shared libraries: /usr/local/lib/libaudiofile.so.0: ELF file OS ABI invalid Well, of course the OS ABI is invalid. The libaudiofile library it's trying to load is for FreeBSD, not Linux. But, if I move the libaudiofile.so.0 file from /usr/local/lib to /usr/local/lib/old, and try to launch gnome-run, everything works fine. So, if the FreeBSD version of the library isn't present, it then looks for the Linux version. Any ideas what's going on or how to fix this? Run /compat/linux/sbin/ldconfig. That should update the path cache of the (linux) dynamic linker. I assume the permissions on the libs already are reasonable. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: NFS - what troubles to expect ?
On Fri, 6 Dec 2002, Josh Brooks wrote: Helol, I have used NFS on sun/solaris systems for many years, and the one big headache that comes to mind is how, if the server is down, and the client has an automount on boot, the client will hang forever (basically forever) trying to mount from the down NFS server. Does this behavior exist in the FreeBSD world ? What other related behavior can I expect if I have about 10 servers that are all _clients_ for an 11th NFS server, and that NFS server goes down ? Will the performance on the clients go down when the NFS server disappears ? Will they too, like solaris systems, hang forever on boot when the nfs server they automount is not present ? It should be configurable. From the mount_nfs(8) man page: By default, mount_nfs keeps retrying until the mount succeeds. This behaviour is intended for filesystems listed in fstab(5) that are critical to the boot process. For non-critical filesystems, the -b and -R flags provide mechanisms to prevent the boot process from hanging if the server is unavailable. Read the rest at http://www.freebsd.org/cgi/man.cgi?query=mount_nfs. If you are using amd(8) to auto mount, options should be tweakable in amd.conf. FreeBSD does not have an exact equivalent to the Solaris automounter. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: ftpd.conf
On Sat, 26 Oct 2002, Ms Carlsson wrote: i do not have one, why waste your time to answer me if you dont even can help me, this is not just a waste of your time it is waste of mine too. so is here anybody serious here, i need a working example of ftpd.conf to the ftp daemon /usr/libexec/ftpd Nope, at least not anybody running FreeBSD. The ftpd.conf file is used by lukempftd (/usr/libexec/lukemftpd), an alternate ftp daemon from NetBSD that has been sort-of-but-not-quite imported into FreeBSD. Configuration of the stock ftpd is described in ftpd(8): it does not use ftpd.conf. The lukem version of ftpd is described in lukemftpd(8). Yes, it is confusing. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: mergemaster question
On Sat, 26 Oct 2002, David Banning wrote: I am just in the process of upgrading from 4.5S to 4.7 and I will be running mergemaster. One thing I notice using mergemaster is that it seems to take such a long time. Is there a quicker way? It wants to replace alot of files, and that's OK for me 95% of the time. Only a few I want to keep or merge. Am I missing something, or is it just a part of upgrading that you must decide, file by file which will be modified or replaced? There will only be a lot of files to update if it has been a long time since you last rebuilt everything and ran mergemaster. You can reduce the number of files to check by copying files you know you have not modified before running mergemaster, e.g.: cp /usr/src/etc/mtree/* /etc/mtree cp /usr/src/etc/defaults/* /etc/defaults cp /usr/src/etc/rc /etc cp /usr/src/etc/rc.{diskless*,i386,network*} /etc I think there are hooks in mergemaster that allow you to script some of this stuff, but I've never gotten around to use them. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: /dev/urandom is randomly cool
On Sat, 5 Oct 2002, Peter Leftwich wrote: I was sorting through my /usr/X11R6/bin/startx text-file and noticed: mcookie=`dd if=/dev/urandom bs=16 count=1 2/dev/null | hexdump -e \\%08x\\` I started playing around with `cat /dev/urandom` and `head -1 /dev/urandom` so my question is... How can I use the head -1 method and change the output into just [A-Za-z0-9] and no spaces or punctuation? And specify length? Why head -1...? /dev/random isn't very line oriented. Oh, well, nevermind. One way of doing what you want is: tr -cd a-zA-Z0-9 /dev/urandom | dd bs=$len count=1 2/dev/null Which will give you $len random bytes from the set a-zA-Z0-9 (it reads a lot more from /dev/urandom than it produces though). Another answer is, as always, use perl :-) $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: /dev/urandom is randomly cool
On Sun, 6 Oct 2002, Fernando Gleiser wrote: On Sun, 6 Oct 2002, Mikko Työläjärvi wrote: tr -cd a-zA-Z0-9 /dev/urandom | dd bs=$len count=1 2/dev/null Which will give you $len random bytes from the set a-zA-Z0-9 (it reads a lot more from /dev/urandom than it produces though). yes, and that is bad :( I know. That is why I mentioned it. You snipped the part where I said that the above is one way of accomplishing the task, as opposed to the only way or the best way, much like slowsort is one way to sort data :) If the characters / and + are added to the set of acceptable output chracters, then the solution is dd the right amount of data and feed to your favourite base64 encoder. As the problem was formulated, you'd need a base62 encoder. It is not good to mess with /dev/[u]random more than what's really needed, because you can exhaust the entropy pool, and that's a Bad Thing. /dev/urandom does not get exhausted, it just gets diluted. Still sub-optimal, but not a total disaster. In your home box, for learning purposes, that's OK, but in a production box which needs a good working prng (for crypto session keys, auth cookies and the like) it is not acceptable to eat all the entropy pool unless you have a very good reason to do so. Agreed. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: /var/spool/uucppublic
On Tue, 1 Oct 2002, Socketd wrote: Hi all The above dir is world writeable, so I just want to know if it is installed as part of the base system and that program/process uses that dir? Can I mount it as no-exec and no-suid? It is part of the system and used by UUCP. As you do not seem to know what uucp is, I'd say it is highly likely that you are not running uucp. Just delete the directory. And keep deleting it after every new upgrade, install or make world, until uucp finally is divorced from the base system, causing grief for all of the three people still alive using uucp... ;-). Hmm.. looks like that will happen in 5.0. Whee. $.02, /Mikko P.S, My first email address was a UUCP address, with a ! and all... To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: man tcsh | col -b | grep -i stderr
On Sun, 29 Sep 2002, Peter Leftwich wrote: I would like to learn more about redirection and stderr but the manpage for tcsh does not have any information apparently about 2 or whatever: Search for redirect... # man tcsh | col -b | grep -i stderr # [no output] Can someone direct me to the proper manpage? `man stderr` did not have the information, nor did `man sh` - please help. Thanks kindly, The format for redirects differ between csh (and derivates) and sh. As I recall, csh can send stderr to stdout, and not much else (you append a to whatever redirector you're using, e.g. or |). I haven't used [t]csh for a very, very long time. To be able to decipher the documentation of redirection in sh, you need to know the numbers of the standard descriptors: 0 = stdin, 1 = stdout, 2 = stderr. $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: Bourne shell redirection of STDOUT
On Sat, 28 Sep 2002, Jimmy Lantz wrote: Hi, I wonder if anyone know a way to redirect the STDOUT directly to a variabel in a shellscript w/o using tempfile. Use backticks ``. I know I can use a tempfile but I'm looking for a way to avoid using a file. in other words Does anyone have an alternative to this? I would like to be able to run this script read-only : #!/bin/sh DIALOG=${DIALOG=/usr/bin/dialog} dialog --inputbox Hitme 8 40 \ 2 /tmp/tempfile myvar=`cat /tmp/tempfile` Aha. You're not redirecting stdout, you want stderr, and backticks will only get stdout (which you don't want, as it is used to present the dialog). Here are two possible solutions: 1) Swap stdout and stderr: myvar=`$DIALOG --inputbox Hitme 8 40 91 12 29` 2) Since dialog will require a tty anyway, let it talk directly to /dev/tty, and redirect stderr to stdout: myvar=`$DIALOG --inputbox Hitme 8 40 21 /dev/tty` $.02, /Mikko To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message
Re: What kind of PPP connection do you have?
On Sat, 14 Sep 2002, Greg 'groggy' Lehey wrote: I'm currently revising my book The Complete FreeBSD. The new edition should be out in the bookstores in a couple of months. Right now I'm revising the chapter on PPP setup, which was written in 1997. Since then, the ISP landscape has changed dramatically, and I suspect that large parts of the chapter are no longer relevant. I'd like your feedback on the following issues: 1. What kind of authentication do you use? Please don't reply if it's PAP or CHAP; I think that's almost universal. I'm just interested if anybody is using anything else (like a UNIX login, for example). Haven't used a login chat script with ppp for at least five years... :) 2. Do you use user PPP or kernel PPP? I'd like answers in each case here, please. I note that other books have decided to describe only one or the other, but currently I think there's a reason for both. User PPP. Hmm... can kernel PPP do PPPoE? /Mikko Mikko Työläjä[EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-questions in the body of the message