From nobody Wed Jan 8 20:20:57 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Installing on Asus A7N266-E mb Date: 08 Jan 2003 20:20:57 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 23 I'm installing 4.6.2 on an Asus A7N266-E. This is the motherboard that has everything integrated -- NIC, video, and sound -- with the NVidia chipset. (According to the CD that came with Windows drivers, it's revision 27.04.) I've installed FreeBSD okay, but it doesn't recognize the integrated sound or network components, only the video. I get several lines like this: pci0: (vendor=0x10de, dev=0x01c3) at 4.0 irq 10 in my dmesg output. The NIC is supposed to be a RealTek RTL 8139, which should be supported by the rl driver in the generic kernel, but it's not recognizing that at all. Is anyone else using this motherboard successfully? Any ideas on how I can get these components to show up and be recognized by FreeBSD? Thanks, -- Aaron From nobody Thu Jan 9 06:44:52 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Installing on Asus A7N266-E mb Date: 09 Jan 2003 06:44:52 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Pasi Savanainen writes: > I quess that you are using GENERIC kernel. GENERIC does not contain > sound devices by default. True; I'd forgotten that the sound devices wouldn't show up at all until I recompiled. But the integrated RealTek RTL 8139 network card is supported by the rl driver, which is already in the GENERIC kernel, yet it's not being recognized. Any ideas on that? Thanks, -- Aaron From nobody Thu Jan 9 14:51:47 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Installing on Asus A7N266-E mb Date: 09 Jan 2003 14:51:47 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 31 > I'm installing 4.6.2 on an Asus A7N266-E. This is the motherboard > that has everything integrated -- NIC, video, and sound -- with the > NVidia chipset. (According to the CD that came with Windows > drivers, it's revision 27.04.) > I've installed FreeBSD okay, but it doesn't recognize the integrated > sound or network components, only the video. I get several lines > like this: To followup, I've now got sound working after compiling a new kernel with: device pcm options PNPBIOS No luck on the network card yet, but I can toss a PCI card in it for now, so that's not a big issue. When XP was on the machine, it reported the NIC as "NVidia nForce MCP" something, not the RealTek that Asus's web site says it has, so who knows. The GeForce2 appears to be recognized ok, but when I try to start X, the system locks up tight. This appears to have been reported several times by others, so I'll try the kernel patch others used and see if it fixes that. -- Aaron From nobody Thu Jan 9 15:04:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Downsize UFS filesystem References: <3e1d9d01$0$203$4d4ebb8e@read.news.de.uu.net> Date: 09 Jan 2003 15:04:12 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 28 Lejf Diecks writes: > my company runs a mailserver on FreeBSD 4.5. Mail is stored in > /var/mail, but the filesystem is to small, especially if users try > to exchange huge mails (10M+). The other posts about resizing/adding partitions are fine, but you could also just move the contents to another partition and create a link. Stop all mail traffic, move /var/mail to /usr/mail (or wherever you want it) with something that will preserve ownership and permissions, and create a symbolic link to it. Something like this: ls -d /usr/mail # make sure /usr/mail doesn't exist already (cd /var && tar cf - mail ) | (cd /usr && tar xpvf - ) cd /var mv mail mail2 # save old copy until done just in case ln -s /usr/mail /var/mail Once you're sure everything's working, remove /var/mail2. This might not be the best solution for an extremely busy server, since it adds the overhead of a symbolic link, but it should be fine for most cases. -- Aaron From nobody Thu Jan 9 15:15:06 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mod_php4: Anyone gotten it to build on 4.7 yet? References: Date: 09 Jan 2003 15:15:05 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 miketemp3@techinxs.com writes: > I'm trying to get an Apache/SSL/PHP4/MySQL/etc. server set-up on a > newly installed 4.7 machine. > > I installed apache13-modssl and mysql323-server. When you install mysql323-server (or maybe it's only when you install mysql323-client, I forget) it creates a startup file at /usr/local/etc/rc.d/mysql-client.sh. All this does is: /sbin/ldconfig -m /usr/local/lib/mysql In other words, it makes the location of the mysql client libraries available to be compiled and linked against. Run that as root, and the mod_php4 port should be able to find mysql. -- Aaron From nobody Thu Jan 9 17:30:32 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Installing on Asus A7N266-E mb Date: 09 Jan 2003 17:30:31 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Yet another followup: After reading , I patched sys/i386/i386/i686_mem.c to disable the i686_mrstoreone routine, recompiled my kernel, and rebooted. The machine still locks up tight as soon as I run startx (after one flicker). It looks like my next move is going to be to upgrade the kernel to the latest STABLE and XFree86 to the latest port, and see if that makes a difference. Any other suggestions in the meantime? Thanks, -- Aaron From nobody Thu Jan 9 19:43:23 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Downsize UFS filesystem Date: 09 Jan 2003 19:43:23 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 11 "John Shafto" writes: > How much overhead can/does a symlink add? Some, I suppose, but unless the server's being pushed to its limits already, I wouldn't think it'd be noticeable. -- Aaron From nobody Fri Jan 10 05:53:39 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mod_php4: Anyone gotten it to build on 4.7 yet? Date: 10 Jan 2003 05:53:39 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 miketemp4@techinxs.com writes: > Well, I tried this and I still get the same error. I have > run both: What's the actual error again? I didn't see it in your original post. Can you post the last few lines that are displayed before it dies? -- Aaron From nobody Fri Jan 10 06:07:34 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: I'd rather ppp not dial when the machine boots References: <0y7T9.4047$bB6.901@sccrnsc02> Date: 10 Jan 2003 06:07:34 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 spaz writes: > How does a person trace what program/process is doing this lookup? First run 'netstat -n -f inet'. This will show you all current IP connctions in and out of your machine. You'll see lines like this: tcp4 0 0 215.137.1.233.1151 62.80.65.68.22 ESTABLISHED That shows my local machine connecting from port 1151 to the SSH port (22) on a remote server. If that isn't enough info to tell me what process is doing it, I'll do: lsof -i TCP:1151 Which returns the command and process number which is bound to that port, along with some other info. The lsof program in in the ports tree in sysutils/lsof. -- Aaron From nobody Thu Jan 16 05:34:10 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mergemaster References: Organization: ESC Date: 16 Jan 2003 05:34:09 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 talon@lpthe.jussieu.fr writes: > I do concur with you. Mergemaster is really something to be > replaced, the sooner the better. Hehe, the first time I used mergemaster I thought it was the coolest thing in FreeBSD yet. But then I've got a mostly Linux background, where there's nothing even that handy for upgrading. My idea for improving mergemaster would be to let me keep a list of files for it to skip, and have an option for it to automatically install the new versions of all others. So I could have, say, /etc/mergemaster.ignore, looking something like this: /etc/master.passwd /etc/group /etc/rc.firewall And then something like 'mergemaster --force --ignore-file=/etc/mergemaster.ignore' would install all the other new files and skip these without asking me about each one. -- Aaron From nobody Thu Jan 16 10:30:49 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mergemaster Organization: ESC Date: 16 Jan 2003 10:30:49 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 talon@lpthe.jussieu.fr writes: > You are right, this is the first point. It is extremely boring to > have to skip a lot of files that mergemester should not touch. You > then end up skipping files that it *should* One thing I've found that helps is to set the PAGER variable to 'tail -30' (on remote systems over slow connections) or 'cat' (on local systems or over fast connections). That way I don't have to page through every diff that's longer than the screen before moving on. I'm going to either 'install' or 'leave for later' based on whether I've customized the file anyway, so I really don't need to see the entire diff at that point. -- Aaron From nobody Thu Jan 16 15:09:42 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mergemaster Organization: ESC Date: 16 Jan 2003 15:09:41 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 jthunder@math.niu.edu (Jeffrey Thunder) writes: > But wouldn't that be unsafe? I'm a freebsd novice and have only done > one upgrade (from 4.4 to 4.7), but distinctly remember having to > truely "merge" /etc/master.passwd and /etc/group, i.e., make changes > to the old ones. Perhaps this particular upgrade was somewhat > unique? I'd still want mergemaster to tell me at the end that it left files X, Y, and Z alone, and keep copies in temproot for me to look at manually. That part should stay the same; I just want to automate the part where I sit there pressing 'i' repeatedly through a bunch of stuff in /etc/mtree or /etc/periodic, hoping I won't doze off and accidentally hit one too many 'i's on master.passwd or something. I've always got a backup, but still... -- Aaron From nobody Sat Jan 18 15:06:22 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Reversing the line order of a file References: Organization: ESC Date: 18 Jan 2003 15:06:21 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 30 Homer Simpson writes: > If you have enough RAM to hold the file in memory, then a simple > perl script will to it: > #reverse.pl > while(<>) { push(@lines, $_); } > print reverse(@lines); No need to mess with the while loop and array: $ cat infile line 1 line 2 line 3 line 4 $ perl -e 'print reverse <>' Organization: ESC Date: 20 Jan 2003 07:39:57 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 michaelrmgreen@yahoo.co.uk (Michael Green) writes: > I've searched the handbook and freebsd.org and the newsgroups but I > still can't understand how ncurses is installed. My objective is to > have a working tic with which to compile new terminfo entries which > MS FoxPro Unix requires. Thanks for any advice. cd /usr/ports/*/ncurses make install See . -- Aaron From nobody Tue Jan 21 10:43:01 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Primer for (potential) FreeBSD converts who are used to Linux? References: Organization: ESC Date: 21 Jan 2003 10:43:00 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 71 Warren Postma writes: > 1. What utilities are there on my system that I as a Linux guy > haven't heard of, but should learn about first? I'd say the two biggest things are the ports tree and the way you update the system and kernel. The ports tree is simply awesome. It has the same handy features of the best Linux software installers, but it compiles everything from source. So programs get compiled against the libraries you have, cutting way down on the type of conflicts you get with rpm-based systems. Upgrading is great, too. On all my systems, a nightly cvs job provides me with the latest source. When a new security patch comes out, or I want to rebuild my kernel for some reason, it's just a matter of running a few commands to upgrade everything to the latest release that I've chosen. One big difference is that FreeBSD documentation tends to be gathered into fewer documents. The Handbook has practically everything you need to know to run a typical workstation or server. When I first started using FreeBSD, I thought the documentation seemed scanty, but that was just because I was used to hunting through hundreds of HOWTOs to find what I needed. > 2. I already noticed that the Ctrl-Alt-F1/F2/... virtual console > thingy works great in FreeBSD. Awesome. Ctrl-A and Ctrl-E do what I > expect. But what keyboard stuff is different than Linux? I haven't really noticed anything. You can use all the same shells and terminals that you have in Linux, so things should work the same. Possibly the biggest difference is that the default shell in FreeBSD is /bin/sh, unlike /bin/bash in Linux, and there are a few differences. /bin/sh doesn't do Emacs-style keyboard controls by default, for example. But it's easy to add that, or to just install bash from the ports tree. > 3. Differences in standard utilities. I assume BSD Tar and BSD Vi are > included instead of GNU-Tar and VIM, but perhaps there are more subtle > things I'm mising. BSD Tar had annoying path name limitations last > time I checked, and lacked support for GNU-Tar extensions, on > religious grounds. Any other similar items I should know of? I've seen discussions of minor differences between the vi programs, but I've never noticed anything in my own use. I'm probably not enough of a power-user to run into the differences. On my FreeBSD systems, the 'du' command has the -d option, which is very handy. My linux systems don't seem to have that. Also, I use 'free' quite a bit on Linux; on FreeBSD you can use 'vmstat' instead, but its output isn't as clear at first. 'top' will tell you the same stuff, though. > 4. Device filesystem differences. I noticed /dev/hd{a,b,c,d} on > (linux) becomes /dev/ad{1,2,3,4} and so on. I like the FreeBSD names > better, so far. I have no idea about any corner cases or unusual > aspects of the new 5.0 devfs. Any primers on the web? The strange part (from a Linux/Windows perspective) is that FreeBSD normally makes a single physical partition (slice) on the disk and then splits that up for its own use. I'm not sure how that works if you've got a dual-boot system and you want the other OS to be able to mount your FreeBSD partitions. Also, FreeBSD doesn't have nearly as much going on in /proc as Linux does. -- Aaron From nobody Tue Jan 21 11:00:32 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advice on installing ncurses Organization: ESC Date: 21 Jan 2003 11:00:32 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 michaelrmgreen@yahoo.co.uk (Michael Green) writes: > Thank you for the reply. I tried: > # cd /usr/ports/devel/ncurses > # make install > I got the error message: > ncurses-5.2 is forbidden: already in the base system. > Any advice on how to proceed? Thanks. You might want to update your ports tree. When I try it, it wants to install ncurses-5.3, and it doesn't say anything about being forbidden. -- Aaron From nobody Tue Jan 21 14:16:17 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: pause a process References: <3e2ce3af$0$230@hades.is.co.za> Organization: ESC Date: 21 Jan 2003 14:16:16 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 bv@wjv.comREMOVE (Bill Vermillion) writes: > I put the following in my /etc/make.conf file: > > FETCH_CMD=/usr/local/bin/wget -c --tries 2 Thanks; that's very cool. I was trying to do this a while back, but by giving the right args to fetch, and didn't have much luck for some reason. It didn't even occur to me to use a different program altogether. -- Aaron From nobody Fri Jan 24 14:24:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: nfs server References: Organization: ESC Date: 24 Jan 2003 14:24:41 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 kevin writes: > i am running a nfs server and ipfilter. is there a way to tell > the nfs server to use just some ports so i could define an ipf rule for > that? AFAIK, nfs uses port 2049 by default. You can also tell it to only listen on a certain IP with the -h option. You'll probably also be running portmap (port 111), mountd (1023), and rpc.statd (udp-1011 and tcp-1022). The easiest way to make sure you're getting them all would be to run 'netstat -a -f inet -n' once after booting with nfs turned off, and once after booting with it on, and then compare to see what's been added. -- Aaron From nobody Fri Jan 24 14:27:21 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: rsync woes.................... References: <3E300815.2040901@flash.net> Organization: ESC Date: 24 Jan 2003 14:27:21 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 Courtney Thomas writes: > I am trying to use rsync to..... mirror my HD to another backup > HD. All local. > The difficulty is that /origRoot = 69506K and /mirrorHDroot = 57096K > after rsyncing. I assume they should be equal [approx]. Where is the > 12500K on /origRoot that is not on /mirrorHDroot ? I haven't tried this on FreeBSD, but I know that on Linux systems there are dynamically-created files in /proc that will basically grow endlessly when you read them, so it's best not to try to back up /proc. You may run into a similar problem if you're backing up /dev. I'd try excluding both of those, and see what happens. -- Aaron From nobody Fri Jan 24 14:31:31 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Multiple DMZs via IPFW help needed... References: <1043492664.360107@news.islandnet.com> Organization: ESC Date: 24 Jan 2003 14:31:31 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 steve@VancouverIslandRobotics.org writes: > Basically, I want to have 2 NICs, one to the world, the other to a > switch with a half-dozen web (and misc.) servers behind hanging on > a switch. By assigning aliases to the incoming NIC, can I not use a > generic redirect rule to forward all traffic for that IP (alias) to > the appropriate box? I'd think that should work, but if you have enough IPs to assign one alias for each box, why not just assign them to the boxes, route to them through your firewall, and forget NAT? Either way your firewalling is the basically the same, but you won't have all the extra hassle of the NAT translations. -- Aaron From nobody Mon Jan 27 08:45:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: libintl.so.4 missing References: <9uf65salnto.fsf@eos.idi.ntnu.no> Organization: ESC Date: 27 Jan 2003 08:45:11 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 12 Oyvin Halfdan Thuv writes: > where can I get it? > > I get this error during compilation of /usr/ports/java/jdk13. 'pkg_info -W' on my system says it's part of the gettext port. -- Aaron From nobody Tue Jan 28 07:07:29 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What? No MC as standard? (Install gripe) References: <20030128074227.72355532.asg@encephalon.de> Organization: ESC Date: 28 Jan 2003 07:07:29 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 23 Mike-UK writes: > No HDD space to install the ports collection That's rough; the ports collection is one of the best features of FreeBSD. I'd make room for it if at all possible. > > "mc" is ugly and in a base installation there is NO need for that. > ??? You like typing then? cd /x cd /y cd /z etc.? <:( I wouldn't say I 'like' typing, but until I can send commands directly from my brain, it's by far the fastest way to move through the filesystem and do useful work. Sure, you can click around through directories pretty quickly with a GUI file manager, but usually I'm cd'ing into a directory because I need to *do something* there, at which point I'll probably have to lean up to the keyboard again anyway. -- Aaron From nobody Tue Jan 28 07:13:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Installing FreeBSD/X on Asus A7N233-VM Organization: ESC Date: 28 Jan 2003 07:13:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 46 I got some help from folks here on getting FreeBSD and X working on my Asus A7N233-VM motherboard, so I thought I'd report back on my success. (This motherboard has all the (NVidia) components built into the motherboard; no PCI or AGP cards at all.) Everything works now except the onboard NIC, so I'm just using a PCI NIC instead. I upgraded to 4.7-STABLE, by doing a cvsup with the RELENG_4 tag and building world. I also rebuilt the kernel, after adding the USER_LDT option to the kernel config file. I also added a couple lines to get the onboard sound working. Here are my additions to the generic kernel: device pcm options PNPBIOS options USER_LDT Rebooting with the new kernel got the sound working right away, but X still froze. So I tried to upgrade X by installing the x11/XFree86-4 port, but it said it was already installed. I discovered that I had to install the parts individually, but installing x11/XFree86-4-*, and especially x11-servers/XFree86-4-Server. That got me to version 4.2.1 of X. Then I installed the nvidia driver from the x11/nvidia-driver port. It installed, but when I went to startx, it brought up a messed-up screen. No crash, though! So I put the default XF86Config back, dearchived the nvidia package, and did a 'make setup' like their docs say. I told it to use its own AGP driver. That made some changes to my XF86Config and /boot/loader.conf that seemed to do the trick, and a reboot and startx brought up a normal X screen. A little fiddling with the Screen section of XF86Config to get the resolution I wanted, and I was all set. My only remaining annoyance is that my AOpen optical wheel-mouse is a little jerky. I've tried it with moused running and /dev/sysmouse, and moused off and /dev/psm0, with no noticeable difference. It's usable, but not as smooth as it should be. -- Aaron From nobody Tue Jan 28 15:59:01 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What? No MC as standard? (Install gripe) Organization: ESC Date: 28 Jan 2003 15:59:01 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 bv@wjv.comREMOVE (Bill Vermillion) writes: > If he sets up cvsup he can refuse many portions of the tree > he'd never need and shrink the amount coniserable with the > advantage of having an always uptodate ports tree if he runs it > from cron. I really have no need for such thinks as Chinese or > Astronomy for example. True. My /usr/local/etc/cvsup/sup/refuse lists all the language-specific subdirectories in the ports tree. I don't know how much space it saves, but when you're cvsup'ing over a modem, every bit helps. -- Aaron From nobody Tue Jan 28 16:11:17 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie working on apache 2.0 install References: <3%wZ9.13164$Xq4.468480@news20.bellglobal.com> Organization: ESC Date: 28 Jan 2003 16:11:17 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "Jim Gray" writes: > The error log states "attempt to serve directory - > /usr/local/www/data" Does that document root sound right? Yes, that's the default. Does that directory exist, and does it contain an 'index.html' file? The default apache config doesn't usually allow directories to be displayed ("Options Indexes"), so if there's no index.html to display, that could generate the error you're getting. -- Aaron From nobody Tue Jan 28 16:30:26 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What? No MC as standard? (Install gripe) References: <20030128074227.72355532.asg@encephalon.de> Organization: ESC Date: 28 Jan 2003 16:30:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 46 Mike-UK writes: > The great beauty of MC is that it is a console FM designed to make > life easier be being primarily keyboard driven. Sure, you can use a > mouse if you miss M$, but its faster than a mouse could possibly be, > and takes a whole lot less keystrokes to get a job done, therefore > reducing the risk of RSI, whuck is kinda handy if you have > arthritis. Here's a job I had to do recently: cd /usr/src find . -type f | xargs grep -i nvidia How many keystrokes would that take with MC? Is it even possible? I'm not being facetious here; I'm asking seriously, as that's an example of a task I do rather often. I can appreciate that a file manager makes certain basic tasks easier, but it also separates you from some powerful shell scripting tools. Heck, I've always got Emacs running because I use Gnus, and Emacs has the dired mode that would let me keystroke through directories and files, but I never use it. The command line is just plain faster. I agree with you than *any* method that uses the keyboard is faster than using the mouse, though. > I'm still trying to get to grips with just how many folks on this NG > don't use MC. You're having me on aren't you? Its a newbie windup > surely? :) I doubt it. I don't know any Unix people who use MC or anything like it. The folks I know fall into two groups. You have the old hands who were doing this stuff before file managers existed, and have never seen a use for one. Then you have the guys who started on Windows, who use a GUI FM like kfm because they haven't gotten comfortable with the command line yet, and clicking around through a GUI is what they're familiar with. But hey, if you like MC, more power to you. -- Aaron From nobody Tue Jan 28 16:33:53 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What? No MC as standard? (Install gripe) References: <20030128074227.72355532.asg@encephalon.de> <20030128170202.1b95e7a7.asg@encephalon.de> Organization: ESC Date: 28 Jan 2003 16:33:53 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Mike-UK writes: > Try Slackware. Install what you want. The newbie custom install > options are magnificent. 30mins to install, then about 10mins to do > basic configurations. Slackware is my favorite Linux. Partly because it hasn't been saddled with a fancy installer (it's the only Linux that would install on my 8MB laptop) and partly because it's the first Linux I used way back when I installed it all from a big stack of floppies. -- Aaron From nobody Tue Jan 28 19:08:01 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie working on apache 2.0 install Organization: ESC Date: 28 Jan 2003 19:08:01 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 DEBO Jurgen writes: > >Yes, that's the default. Does that directory exist, and does it > >contain an 'index.html' file? The default apache config doesn't > >usually allow directories to be displayed ("Options Indexes"), so > >if there's no index.html to display, that could generate the error > >you're getting. > Aaron, that's the default, but isn't a directory /var/www more > suitable with the needed permissions apache/webdaemon ? You lost me; I don't know what permissions have to do with the path you use. I'd say it's just personal preference: use whatever path you want, and change it in httpd.conf if you don't want the default. Ownership of the directory doesn't matter to apache, as long as the permissions allow it to read the files. If you're using CGI with suexec, that's another matter, but that's not under the same directory anyway. -- Aaron From nobody Tue Jan 28 19:11:39 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: avifile won't make References: <3E3711C4.87C65049@pacbell.net> Organization: ESC Date: 28 Jan 2003 19:11:39 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Jam writes: > I tried to install avifile from /usr/ports/graphics/avifile. > The make file wanted to retrieve avifile-0.7.11-20020711.tgz from > the ftp.freebsd.org site but the file wasn't there. It looks like you need to update your ports tree. On my system (updated nightly) avifile is in /usr/ports/multimedia/avifile, and it's version 0.7.22-20021129. -- Aaron From nobody Wed Jan 29 06:55:14 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie working on apache 2.0 install Organization: ESC Date: 29 Jan 2003 06:55:12 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 32 Jim Gray writes: > I checked all that and the directory does exist. It contains all the > default apache files. Yes, but does it contain one called 'index.html'? My default install doesn't; it just has a bunch of language specific versions. Copy one of those to 'index.html', and see if that helps. > The one odd thing is that the name of the directory > should be /usr/local/www/data however KDE lists it as data-dist I noticed that too. Make sure your httpd.conf has 'Options FollowSymLinks' set. My installation is 1.3, not 2.0, but I think that should be the same. I'll install 2.0 today, and see if that makes a difference. > I have tried changing the root but I get an error. It still is "501 > method not implemented" and I am using the default Apache config > file. I'm not sure what you mean by 'changing the root', if you haven't changed the default httpd.conf. Make sure an index.html exists in your DocumentRoot, and let us know if that works, and what shows up in /var/log/httpd-error.log if it doesn't. Also, try a URL like this: http://localhost/index.html.en -- Aaron From nobody Wed Jan 29 07:06:32 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What? No MC as standard? (Install gripe) Organization: ESC Date: 29 Jan 2003 07:06:32 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 38 bv@wjv.comREMOVE (Bill Vermillion) writes: > >I'm not being facetious here; I'm asking seriously, as that's an > >example of a task I do rather often. > I was. > [being facetious that is] In retrospect, I probably was too. It's so hard to tell sometimes. > >I doubt it. I don't know any Unix people who use MC or anything > >like it. The folks I know fall into two groups. You have the old > >hands who were doing this stuff before file managers existed, and > >have never seen a use for one. Then you have the guys who started > >on Windows, who use a GUI FM like kfm because they haven't gotten > >comfortable with the command line yet, and clicking around through > >a GUI is what they're familiar with. > I'll go along with that. I've put it on to try it - but it took so > many more key presses to do anthing. But for a new person to Unix > it may be what it take to get them going and then they can move on. If I could move one single thing from the ports tree to the default install, it'd be 'lsof.' I hardly ever use it except working on an urgent problem, and that's not when I want to take the time to install a port. Now it's the first port I install as soon as the system is online. But, if the maintainers added everyone's favorite program to the default install, we'd have -- Redhat. -- Aaron From nobody Wed Jan 29 07:15:31 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Multi Boot with XP References: <3e37aebc$0$2571$afc38c87@news.easynet.fr> <3e37b941$0$2567$afc38c87@news.easynet.fr> Organization: ESC Date: 29 Jan 2003 07:15:31 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 27 Melwyn Da'alimar writes: > Is the one provided with FreeBSD a graphical one ? Or text based one? It's very text-based. I had XP on the first partition, then installed FreeBSD on the second, telling it to write the MBR. When it boots I get a menu: F1: ? F2: FreeBSD I press the F-key for the one I want, or after several seconds it defaults to the one that was booted last. > The hardest part to use FreeBSD will be to make some room on my hard > drives to create a freeBSD partitions, all my drives are full of > stuff :) If your partitions cover your drives but you have free space, you can resize them. I downloaded a utility called bootitng; it was a pain to install and use, and it took forever, but it did the job of resizing my XP partition to make room for FreeBSD. -- Aaron From nobody Wed Jan 29 15:34:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: I jumped the Windows ship...and am drowning! References: Organization: ESC Date: 29 Jan 2003 15:34:23 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 109 "SV" writes: > First, where can I find some good docs explaining basic things like > how to install software packages, a general command list for the > command line, just a lot of the basics. As someone else suggested, start with the handbook. You might also want to pick up some sort of 'Unix for Dummies' book from your library or bookstore. A lot of the day-to-day commands you'll use for things like manipulating the filesystem are the same across Unix systems, so the book doesn't have to be specifically for FreeBSD, although that'd be nice. > Can I get the command prompt to display the current working path > similar to DOS? I find that I get lost while navigating my > directories. Every shell handles this differently, but it usually involves setting something in your shell's startup file. On systems where I use bash, I have this in my .bash_profile: PS1=$'($PWD/)\n`hostname -s` \\$ ' That gives me a two-line prompt, with my current path in the first line and the hostname in the second. Under zsh, I do this in my .zshrc: precmd () {print -Pn "\e]0;%n@%m: %~\a"} That puts the hostname in my prompt, and puts my username, the hostname, and my current directory in my xterm's title bar. So you'll have to see your shell's man page or other docs to see how to change your prompt. I'm not sure that the default 'sh' shell on FreeBSD can even do it. In the meantime, you can always use the 'pwd' (Present Working Directory) command to see where you are if you get lost. Or just type 'cd' to go directly to your home directory. > What is bash? I hear the name thrown around, but have no idea what > it is. Bash is a descendant of the bourne shell 'sh', commonly used as the default shell on Linux systems. If you've installed the ports tree (and you should) you can install bash by going to /usr/ports/shells/bash2 and running 'make install'. I don't use bash under FreeBSD, simply because I'm already using it on Linux systems, so I like to try other shells like zsh on FreeBSD. > My last question...for now, pertains to GNOME. After the install of > FreeBSD, and the install/configuration of XFree86, I installed > GNOME. What happens is that after logging in, and typing 'startx', X > comes up but I have no control of the mouse(even though it worked > perfectly fine in the graphical X setup), and I get a bunch of > errors that certain functions couldn't start, and a big one that > says Nautilus can't start because of an inability of Bonobo to find > the factory. If you just installed this system and starting over wouldn't lose any important data, you might want to go that route. Now that you've been through it once, you might have a better idea what's going on. When you get to the mouse setup, if you have a PS/2 mouse, use the Auto settings if possible and see if that works. Getting X setup correctly is often the biggest challenge of installing any *BSD or Linux system. > One thing I noticed was that the option just said GNOME desktop I > believe, whereas the documentation I was looking at [1] showed an > option for GNOME + Enlightenment. Does the new install not include > Enlightenment? Am I having problems because I don't have a window > manager, or is there something else wrong here? If you did a typical install with X, you'll have several window managers installed, but it'll have a certain one set as the default in /usr/X11R6/lib/X11/xinit/xinitrc. You can override that by creating an .xinitrc file in your home directory. This file will be executed when you run startx. Mine just looks like this: #!/bin/sh xset +dpms exec wmaker The second line turns on DPMS for my monitor, and the third one starts the WindowMaker window manager. You could replace that with twm, mwm, or whatever other window manager you want to use. Note: If you want to use WindowMaker, run the command "/usr/X11R6/bin/wmaker.inst" from the command line first to set things up. > I also got a message that since I wasn't connected to the internet, > GNOME wouldn't work correctly, is this only a first time thing, or > must I always have an internet connection to use GNOME? I really don't know what it's doing here, but you shouldn't normally need a connection to use it. Maybe it's trying to update itself or something. > Also, how the heck do I get out of X and back to the command line? Ctrl-Alt-Backspace should kill X, but most window managers offer a gentler way to do it. Try clicking on the desktop with your various mouse buttons and see if one brings up a menu with an 'exit' option somewhere on it. -- Aaron From nobody Wed Jan 29 19:45:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: I jumped the Windows ship...and am drowning! References: Organization: ESC Date: 29 Jan 2003 19:45:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 44 "SV" writes: > The mouse starts out in the middle of the screen, and then as soon > as I move it, it snaps to the upper left corner of the screen, and > no matter what I do I can't move it anywhere. This is what I need to > fix? It's a plain old MS PS/2 mouse with scroll wheel. "moused" > shows as starting on FreeBSD boot, and doesn't appear to show any > errors or problems. moused_enable is set to YES. What I don't get is > why it works in the FreeBSD mouse configuration screen, and the > XFree86 screen without a problem, but not when X is actually loaded. That's because the X installer uses a very basic configuration so that it'll work on lots of hardware, while when you start X normally, it's using the specific config that the installer created. Edit /etc/X11/XF86Config and look for an "InputDevice" section for your mouse. Since you have moused running, the section should look something like this (probably with comments interspersed): Section "InputDevice" Identifier "Mouse1" Driver "mouse" Option "Protocol" "Auto" Option "Device" "/dev/sysmouse" Option "ZAxisMapping" "4 5" Option "Buttons" "5" EndSection Don't change the Identifier; that's just used to reference this section later. If setting the Protocol to Auto doesn't work, try "PS/2". Setting the Device to /dev/sysmouse tells it to use moused. The ZAxisMapping and Buttons options are only necessary to get the wheel(s) to work, but those shouldn't keep the basic mouse functions from working. Whenever you make changes to this file, you'll need to quit X and restart it. -- Aaron From nobody Wed Jan 29 19:48:15 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: avifile won't make Organization: ESC Date: 29 Jan 2003 19:48:15 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 Jam writes: > Thanks. I followed your instructions and updated the ports. > > Now when I try to make avifile, it says it can't find > > avifile-07.22-20021129.tgz Same here. It looks like the developers of avifile have a new version out, and have deleted the version that the port tries to get. If you look in the Makefile, you can see where to get it, and go get the latest version and install it from source yourself. -- Aaron From nobody Wed Jan 29 19:53:35 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: sendmail and make world References: Organization: ESC Date: 29 Jan 2003 19:53:35 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 "Peter Wu" writes: > Say, I upgrade to the latest sendmail, whose version is bigger than > the one that comes with 4.7. Then, I do a make world. > My question is whether make world will restore sendmail to the > version when it ships or it keeps the latest sendmail I have > upgraded. See 'man make.conf' for the NO_SENDMAIL variable. Then install qmail. :-) -- Aaron From nobody Thu Jan 30 05:44:26 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: sendmail and make world Organization: ESC Date: 30 Jan 2003 05:44:26 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 "Peter Wu" writes: > I know of that option in make.conf but I want to continue to use > sendmail, at least for now. :) > Since sendmail needs patches relatively more frequently, I wonder > how I can keep patched to be more secure on a FreeBSD system. Well, if you install the latest version from source, and set NO_SENDMAIL in make.conf, you'll have to install your own patches and upgrades when they come out for your version. I suppose there's probably some sort of mailing list for sendmail security notices, so you could join that to hear about them as soon as possible. Any time you install something that's outside the base distribution, you're going to have to keep up on security issues yourself. -- Aaron From nobody Thu Jan 30 06:04:32 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Remote printing problem References: Organization: ESC Date: 30 Jan 2003 06:04:31 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 "Boris H." writes: > This is my /etc/printcap: > lp:\ > :sh:\ > :lp=:rm=*my printserver*:rp=lp2:\ > :sd=/var/spool/lpd/:lf=/var/log/lpd-errs/:\ > :mx#0: I just see one problem there: I'm betting /var/log/lpd-errs is a file, not a directory, so you don't want that trailing slash. You don't need it on the 'sd' value either. I don't suppose that's what's keeping it from working, though. -- Aaron From nobody Fri Jan 31 06:05:24 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Colored shells References: <3e3a3b49$0$2554$afc38c87@news.easynet.fr> <3e3a4018$0$2579$afc38c87@news.easynet.fr> <3e3a415c_3@corp.newsgroups.com> <3e3a433f$0$2576$afc38c87@news.easynet.fr> <3e3a45e9_3@corp.newsgroups.com> <3e3a4afa$0$2567$afc38c87@news.easynet.fr> Organization: ESC Date: 31 Jan 2003 06:05:22 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 Melwyn Da'alimar writes: > And both have the path displayed on screen ? You know, the same > thing as in DOS when we see C:>program files\ Sure. In csh/tcsh, do this in your startup files: set prompt="%~ %# " And in bash, do this: export PS1="\w \$ " Most shells have a bunch of stuff you can do with the prompt. See the 'PROMPTING' section in 'man bash', or the section on the 'prompt' variable in 'man csh'. -- Aaron From nobody Fri Jan 31 06:08:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Colored shells References: <3e3a3b49$0$2554$afc38c87@news.easynet.fr> <3e3a4018$0$2579$afc38c87@news.easynet.fr> Organization: ESC Date: 31 Jan 2003 06:08:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 Melwyn Da'alimar writes: > You mean there's no way to have colors without putting extensions to > ls each time we want to use it ? According to 'man ls', setting the environment variable CLICOLOR does the same thing as giving ls the -G option. So just set CLICOLOR to 'yes' or something in your shell's startup file, and you won't have to worry about what options you give ls. If you're running it in an xterm, you may also need to set TERM to 'xterm-color' to get xterm to display the colors. -- Aaron From nobody Fri Jan 31 06:12:56 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: How to setup an usb mouse ? References: <3e396553$0$29498$afc38c87@news.easynet.fr> <1oqb1b.d06.ln@simonbarner.myfqdn.de> <3e3a2630$0$2556$afc38c87@news.easynet.fr> <3e3a31a1$0$29516$afc38c87@news.easynet.fr> Organization: ESC Date: 31 Jan 2003 06:12:56 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 Melwyn Da'alimar writes: > Thanks :) Mmmm without mouse I cannot use XFree, so I cannot launch > an html page, isnt it ? cd /usr/ports/www/lynx make install distclean lynx /usr/share/doc/handbook/index.html -- Aaron From nobody Fri Jan 31 06:17:23 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Using pkg_add References: <3e3a3de0$0$29498$afc38c87@news.easynet.fr> <3e3a439c@duster.adelaide.on.net> <3e3a4b39$0$2567$afc38c87@news.easynet.fr> Organization: ESC Date: 31 Jan 2003 06:17:23 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 Melwyn Da'alimar writes: > Thanks, so to install those stuff I have to unzip and untar this file > somewhere, then read the instructions on how to install it ? Yes. Generally that means reading the README and INSTALL files in the package after you unpack it, and any others they may refer to. But unless you need version 1.3 of mozilla for some reason, you'd be better off doing: cd /usr/ports/www/mozilla make install clean Right now that'll build and install mozilla 1.2.1 (and anything it depends on). -- Aaron From nobody Fri Jan 31 06:50:58 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Colored shells References: <3e3a3b49$0$2554$afc38c87@news.easynet.fr> <3e3a4018$0$2579$afc38c87@news.easynet.fr> <3e3a415c_3@corp.newsgroups.com> <3e3a433f$0$2576$afc38c87@news.easynet.fr> <3e3a45e9_3@corp.newsgroups.com> <3e3a4afa$0$2567$afc38c87@news.easynet.fr> <3e3a57e9$0$2556$afc38c87@news.easynet.fr> Organization: ESC Date: 31 Jan 2003 06:50:58 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 23 "Melwyn Da'alimar" writes: > How comes as a standard user I cannot see config files in my home > directory, but as a super user I can ? :) You know, that's a darned good question. I always assumed that root's .cshrc was aliasing 'ls' to 'ls -a', but I just checked and there's no alias set for 'ls'. So I checked 'man ls', and found this: -A List all entries except for . and ... Always set for the super-user. So it appears that ls does it internally. If you want your non-root user to do the same thing, do this in your shell's startup file: alias ls='ls -A' -- Aaron From nobody Fri Jan 31 17:06:02 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: changing IP References: <6%C_9.667$HG6.270170@newssrv26.news.prodigy.com> Organization: ESC Date: 31 Jan 2003 17:06:01 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "dsavitsk" writes: > the only changes i made to freebsd was to change (in /etc/rc.conf) > ifconfig_dc0="inet 192.168.0.25 netmask 255.255.255.255" to > ifconfig_dc0="inet 192.168.1.25 netmask 255.255.255.255" Netmasks should be 255.255.255.0 (probably, unless you're playing with subnets). With these netmasks, your systems don't know how to find 192.168.1.1, or any other system except themselves for that matter. -- Aaron From nobody Sun Feb 2 15:30:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Why is Running Out Of Swap So Dramatic? References: <87u1fmq0z7.fsf@pele.r.caley.org.uk> <20030202162116.1d6b1915.steveo@eircom.net> <87vg02ocaa.fsf@pele.r.caley.org.uk> <20030202184403.740ca429.steveo@eircom.net> Organization: ESC Date: 02 Feb 2003 15:30:04 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 Steve O'Hara-Smith writes: > Oh dear - I get that sort of effect for thermal shutdown on > AMD chips. I've got an old K6-233 system that's started doing that lately too. Every few days or so it freezes up for 30-60 seconds, and then reboots. There's never anything in the logs, and I've swapped all the hardware except the motherboard and the CPU, so I know that's the problem. It's just a test system, but it's still annoying. This morning when it did it the fsck process trashed /sbin/fsck. So it couldn't even finish checking the other partitions. I had to mount them read-only, bring up the network, and copy /sbin/fsck from another system. I think I'll look into making / read-only; there's really no need for a crash to be messing up programs on / like that. -- Aaron From nobody Sun Feb 2 15:44:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: gmake ? References: <3E3D364F.2010207@plazaearth.com> <3E3D4926.1020708@plazaearth.com> <3E3D5F87.8090207@plazaearth.com> <3E3D6DE1.50801@plazaearth.com> Organization: ESC Date: 02 Feb 2003 15:44:11 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 philo writes: > at least if i type mc > > or ./mc > i just get command not found Do 'echo $PATH' and make sure it has /usr/local/bin and /usr/X11R6/bin in it. The mc binary will most likely be installed to one of those. Try it with full paths: /usr/local/bin/mc /usr/X11R6/bin/mc and see if one of those works. -- Aaron From nobody Sun Feb 2 15:47:29 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: perl References: Organization: ESC Date: 02 Feb 2003 15:47:29 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 27 "Pons" writes: > On a FreeBSD > #perl -MCPAN -e shell; > ..... > Are you ready for manual configuration? [yes] no > Prototype mismatch: sub CPAN::FirstTime::prompt ($;$) vs none at > /usr/libdata/perl/5.00503/CPAN/FirstTime.pm line 85, chunk 1. Sounds like your CPAN::FirstTime module is screwed up. I'd go download the CPAN module and install it by hand, and see if that fixes it. That probably goes something like this: download CPAN-something.tgz tar -xzf CPAN-something.tgz cd CPAN-something perl Makefile.PL make make test make install -- Aaron From nobody Sun Feb 2 16:52:57 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Why is Running Out Of Swap So Dramatic? Organization: ESC Date: 02 Feb 2003 16:52:57 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 Brent Busby writes: > I had problems similar to this on my machine which turned out to be > nothing more than processor seating in its socket. Reseating the > CPU fixed everything. No guarantees of course that it has anything > to do with what's happening to you. Thanks; it's worth a try. -- Aaron From nobody Tue Feb 4 10:30:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Simple mailing-lists References: Organization: ESC Date: 04 Feb 2003 10:30:11 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Leo writes: > Okay... In fact, I find Majordomo an heavy solution, that's why I > didn't want to manage my lists with it. Is there something a bit > lighter than Majordomo ? I've being talk about Mailman, any > experiences ? I like ezmlm, but I think it requires that you use qmail for your mail server. -- Aaron From nobody Tue Feb 4 13:01:05 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Simple mailing-lists Organization: ESC Date: 04 Feb 2003 13:01:04 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 talon@lpthe.jussieu.fr writes: > which may require recursively you ditch syslog for a djb package, I'd like to ditch syslog, but too many other things use it. I use multilog wherever it's convenient. > and so on at infinity. Perhaps this is the way to get a GNU/BSD/DJB > system? I happen to like and use many of DJB's programs on both FreeBSD and Linux, but your tastes may differ. The original poster asked what mailing list managers people liked, so I told him. -- Aaron From nobody Wed Feb 5 05:07:55 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: cron emails References: <3e40b49b$0$231@hades.is.co.za> Organization: ESC Date: 05 Feb 2003 05:07:54 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "ju" writes: > How can I stop cron emailing the output the the user? I'm running a > cron every minute and it fulls my /var partition sometimes, not > good, thanks If you want to prevent a single cron task from mailing, just redirect its output to /dev/null. Better yet, if possible, alter the task itself so it doesn't output anything when it succeeds. -- Aaron From nobody Wed Feb 5 16:51:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: A few questions..... References: Organization: ESC Date: 05 Feb 2003 16:51:02 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 65 "_Jung" writes: > 1) Apparently, those two apps were not included in the distribution > package that came along with FreeBSD installation CD, at least the > first CD. By the way, what are these extra CD's are for? Do they > contain more application packages? I believe so. But you can install both Apache and PostgreSQL from the ports tree. You did install the ports tree, right? If not, go back and do it now; you'll be glad you did. Then go here: > 2) I didn't install X Windows because I thought that wouldn't be > necessary. However, soon I found myself having to come up to my > room to use my Win98 computer to do article search and application > package downloading because without X I can't have a web browser. > Quite inconvenient. Was I wrong not to have X Windows installed on > the FreeBSD box, even though eventually I am going to use it mainly > web and DB servers? I usually end up installing a big chunk of X on servers, even if they'll never run an X app, just because I end up installing utilities that require various pieces of X. For example, I'll want the ImageMagick package for converting between graphic formats, and that'll depend on xpm, or something like that. So now I just install X from the beginning. But you don't have to have X to browse the web; there are a handful of text browsers in the ports tree, such as lynx, links, and w3m. If you do want to install X, you can do it from the ports tree, starting in /usr/ports/x11/XFree86-4. > 3) What will be the easiest way of sharing files between my FreeBSD > box and other Win9X computers in my house? Is this where NFS comes > in to play? It could, but Samba would be simpler, since your Win boxes already speak SMB. It should be in the handbook. NFS would require running third-party software on your Win boxes, and that's best avoided whenever possible. NFS is best used when sharing mounts between Unix-ish systems. > 5) This is a bit premature question. In the future, when I add > another FreeBSD box on to my home network, do I simply install > FreeBSD on it as if it is another stand-alone UNIX box? It doens't > sound right to have two root accounts... Is there a way of > installing FreeBSD as "add in" box for the exisiting network? If you'd like to share user info across multiple Unix machines, check into YP/NIS. I haven't used YP in years, but it's part of the base FreeBSD install, so I'm sure the handbook explains it. You might still want to keep a separate root account (and password) on each system, though, so that if someone breaks into your gateway system he doesn't have easy access to all the others. A tip: If you don't use YP, then when you create users on the second machine, make sure they have the same UIDs as on the first machine. That way if you want to mount their home directories from one machine to the other via NFS, their permissions will match automatically. -- Aaron From nobody Fri Feb 7 06:01:10 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Web Server and Web Browser References: Organization: ESC Date: 07 Feb 2003 06:01:10 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 "_Jung" writes: > For example, the very first part of the manual talks about > configuring for compilation. It keeps saying to call ./configure > in order to specify some options. However, I > couldn't. Since I didn't extract files from the tgz file, but > simply called "make install" from the ports directory, I didn't have > the executable named "configure." (At least I couldn't find it.) > Is manual installation necessary if I want to configure some > compilation options? No, you can add options to configure in the Makefile. Look for the CONFIGURE_ARGS variable. The contents of that variable will be passed to the configure script. -- Aaron From nobody Fri Feb 7 14:14:46 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: cable gateway issues References: Organization: ESC Date: 07 Feb 2003 14:14:46 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 "BSD Nebwie" writes: > I do an ipfw list and it gives me last two lines are > 65000 allow ip from any to any > 65535 deny ip from any to any > I don't see 65535 in the config file. Where is the firewall drawing > that from? The kernel sticks that one on the end by default if you don't build it with the IPFIREWALL_DEFAULT_TO_ACCEPT option. But rule 65000 matches everything, so you'll never have a packet reach rule 65535 anyway. -- Aaron From nobody Sat Feb 8 06:14:00 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Make readmes error References: <2083176d.0301210832.1aa5222c@posting.google.com> Organization: ESC Date: 08 Feb 2003 06:14:00 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Kris Kennaway writes: > You can't; this isn't supported. If you want to use targets like > 'make readmes' or 'make index' then you need to have the full ports > collection. I don't think so, unless this has changed recently. I have all the non-English language directories in my refuse file, and I've run 'make readmes' without any trouble. I also had to physically delete those directories first. If you don't do that, 'make readmes' tries to recurse through them and runs into trouble. -- Aaron From nobody Sat Feb 8 06:19:30 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: freeBSD 4.7 References: Organization: ESC Date: 08 Feb 2003 06:19:30 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 36 "Philo" writes: > now i've done a fresh installation of 4.7 and have noticed > somewhat decreased performance (on the same machine) > > amd-1000 256megs ram > my feeling is that an amd-1000 should be plenty good for BSD Absolutely. Until recently, I ran FreeBSD (and before that, Linux) on an AMD K6-233 with 96 MB of RAM. Typically I'd be running X with Windowmaker, XEmacs, 6-10 xterms, and Galeon or Netscape. Yes, it was a little pokey, but not drastically so. > the one thing i did differently was to setup my new installation > as one large partition / I wouldn't expect that to cause a noticeable slowdown. After all, even when you divide FreeBSD up into the default partitions, they're still within one physical slice on the disk, unless you make it otherwise. Dividing into multiple partitions is mostly done for reliability and security, not speed. > before i get too far > would it be better to just go with the default partitioning scheme > as i did with bsd4.4 > > or is the problem with kde perhaps??/ Kde is definitely a resource hog. You could try a lighter window manager like fvwm2 or windowmaker and see how that feels. -- Aaron From nobody Mon Feb 10 14:44:34 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advice re error : Shared object "libmytinfo.so.2" not found References: Organization: ESC Date: 10 Feb 2003 14:44:34 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 michaelrmgreen@yahoo.co.uk (Michael Green) writes: > > I am running a shell script I was sent which gives this error: > > /usr/libexec.ld-elf.so.1: Shared object "libmytinfo.so.2" not found > > The line causing this look like this: > > tic fansi.src This is the terminfo compiler. > Can anyone give me a pointer to this problem, what is > libmytinfo.so.2? On my system it exists in /usr/lib/compat, so I assume it's part of the linux compatibility system. Did you install that when you installed FreeBSD? What does 'which tic' return? -- Aaron From nobody Mon Feb 10 15:01:08 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: kernel config file References: Organization: ESC Date: 10 Feb 2003 15:01:08 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 11 polux writes: > Where can I find the config file that was used to make my current > kernel ?? 'uname -v' will give you the path to the file. -- Aaron From nobody Mon Feb 10 19:15:08 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: kernel config file Organization: ESC Date: 10 Feb 2003 19:15:08 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 23 Desmond Coughlan writes: > Le 10 Feb 2003 15:01:08 -0600, Aaron Baugher wrote: > > 'uname -v' will give you the path to the file. > Um, are you sure ? Oops, you're right; the paths are similar enough that I didn't notice the differences. Still, the basename of the path should match the name of the config file, as far as I know. Mine returns /usr/obj/usr/src/sys/cail, and my config file is /usr/src/sys/i386/conf/cail. Unless you're doing something unusual, the config file should be in that directory. How about this: echo -n /usr/src/sys/i386/conf/; basename `uname -v | cut -d ':' -f 5` :-) -- Aaron From nobody Tue Feb 11 10:36:00 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: X launching permissions References: Organization: ESC Date: 11 Feb 2003 10:35:58 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 49 Fritz writes: > BTW.. Is XDM a better way to go.. From what I read I think I > understand both XDM an XFree86 can exist at the same time?? Xdm is the X Display Manager, a part of XFree86. As the man page says: "Xdm manages a collection of X displays, which may be on the local host or remote servers." So its primary purpose is to manage multiple displays, whether on the local machine or from remote machines. Xdm comes in handy if you have other lightweight X servers on your network, like disk-less X-terminals or Windows machines running a third-party X server, since it provides the users on those machines with a direct login to the main X system, where all their programs will run without needing so much as a window manager on their local machine. If your system will only have one person at a time logged in and running an X display, and you won't be displaying X from it to other machines, I don't think you have any particular need for xdm, so there's no point in wasting a little CPU and RAM on it. In that case you should be just fine with startx. [X terminology can be confusing for newbies because of the meanings of 'client' and 'server.' We get used to the idea that the client is the program running locally and the server is something remote, but in X that's usually just the opposite of what happens. In X, a 'server' is a program that controls the local display and makes itself available for clients to connect to. A 'client' is a program that connects to an X server, asks it to display things, and gets user input back. (From a technology standpoint it makes perfect sense, but from a human user standpoint it takes a little getting used to.) So the 'server' will almost always be running on your local machine, while your 'clients' may be running anywhere. Documentation on X, including the man pages, will make much more sense once you understand these terms.] One more thing: When you login through xdm, X runs the ~/.xsession file to start your display environment, so that's where you want to put anything you want to happen whenever you login with X, like starting programs, doing xset commands, and so on. When you use startx, the file ~/.xinitrc is run instead. I usually just create one of these and then do a symbolic link to it from the other one, to keep them the same. -- Aaron From nobody Tue Feb 11 10:40:49 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Is it dangerous to make ports from a higher release? References: <163i4vcfm6pv3mbng20t7bum7nue61kagg@4ax.com> Organization: ESC Date: 11 Feb 2003 10:40:47 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 Yin Ming writes: > I'm using a 4.4 version (maybe not, but at least it's not the newest), > I wander whether it will fail in building a ports downloaded from a > higher version release. And I would not like to download large files > to update my whole system, Can I make ports from the release of my > system's? There's only one ports tree. Ports are updated individually as their maintainers see fit. I would guess that there are many current ports that would compile just fine on a 4.4 system; the only way to find out is to try them. I certainly wouldn't call it 'dangerous'; if it fails, it won't install. If you run into trouble (or even if you don't), you should upgrade your base system and kernel. For that, see the Handbook section on building world. Since you clearly aren't concerned about having the latest cutting-edge stuff, I'd suggest tracking RELENG_4_7 in your cvsup file when you get to that point. -- Aaron From nobody Tue Feb 11 10:44:55 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Release 5.0 (root@hollin.btc.adaptec.com:/usr/obj/usr/src/sys/GENERIC) References: Organization: ESC Date: 11 Feb 2003 10:44:54 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 brent@dbprograms.com (Brent Bolin) writes: > Why does uname show this?. This is from an ISO that was downloaded > and installed. Generic kernel(adaptec.com). Think this is from the > FBSD ftp server but don't recall for sure. It's telling you who compiled the kernel, and where in the compile tree it was located at the time. When you compile your own kernel, your own $USER@$HOSTNAME will show up there, along with the path ending in whatever you named your kernel config file. *Why* does it show it? I suppose because it wants to. -- Aaron From nobody Tue Feb 11 18:23:54 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Release 5.0 (root@hollin.btc.adaptec.com:/usr/obj/usr/src/sys/GENERIC) Organization: ESC Date: 11 Feb 2003 18:23:53 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 brent@dbprograms.com (Brent Bolin) writes: > My question was more along the lines of why would the ISO image > downloaded indicate this usr@node.domain.name . > What does adaptec.com have to do with FBSD?. Since the GENERIC kernel in your version was compiled there, I assume someone there is involved in creating the distribution. You could always write to that email address and ask. -- Aaron From nobody Wed Feb 12 15:27:10 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: I might as well be running KDE! (Mozilla - pig) References: <3e4a9a71_1@corp.newsgroups.com> Organization: ESC Date: 12 Feb 2003 15:27:10 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 Donn Miller writes: > As you can see, it's using upwards of 100 MB RAM! With memory usage > like that, I could probably be running KDE and Konqueror, instead of > Mozilla and WindowMaker. Also, it is using over 13% CPU time as I'm > composing this message. Or you could run Windowmaker and Konqueror. You don't have to be using KDE as your window manager to run KDE apps. I run Windowmaker, and mostly use Galeon as my browser. It uses the mozilla rendering engine, but seems lighter and faster and has some other features I like. Right now it's using 60MB of RAM, but I've also got 4 tabs open and it's been running for a few days. -- Aaron From nobody Fri Feb 14 09:51:58 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: sysctl values References: <1045227935.645268@sestra.vokar.com> Organization: ESC Date: 14 Feb 2003 09:51:58 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 "Vitali Malicky" writes: > we all know how to get kernel variable values with sysctl > > # sysctl -n kern.boottime > # { sec = 1044442100, usec = 881341 } Wed Feb 5 12:48:20 2003 > ok! but now I want to get only that one value i.e. 1044442100 > without shell coding, and with as little perl coding as possible. As long as you know the value you're looking for will be the first string of digits: sysctl -n kern.boottime | perl -ne 'print /(\d+)/' -- Aaron From nobody Sun Feb 16 15:47:58 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Debian To FreeBSD help References: Organization: ESC Date: 16 Feb 2003 15:47:57 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 "Ean Kingston,,,(416) 423-3770" writes: > Add BASH from the ports and use it as your user shell, things will > work the way you want (bash is what debian uses). You should leave > the root shell alone though. Remember, during normal operation you > can login as yourself and use the su command to become root without > losing your shell. Nope. When you su to root, you get root's shell: cail% echo $SHELL /bin/sh cail% su Password: cail# echo $SHELL /bin/csh If I wanted to use a shell like bash as root, I'd put something like this in root's .profile (or .login): if [ -x /usr/local/bin/bash ] ; then exec /usr/local/bin/bash fi That way, if you're coming up in single-user mode and that filesystem isn't mounted, bash won't try to start and you'll be fine. You just don't want to make root's default shell something that might not always be available, especially when there are problems. > # For PnP/PCI sound cards > device pcm Also, if your sound card is integrated into the motherboard, you might need to add: options PNPBIOS -- Aaron From nobody Wed Feb 19 14:26:32 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FFS rant References: <2v775v036hm94npte0h7da15rc80fql8vi@4ax.com> Organization: ESC Date: 19 Feb 2003 14:26:32 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 36 bv@wjv.comREMOVE (Bill Vermillion) writes: > Harvey Gage wrote: > >I myself have experienced power outages that produced random walks > >across the filesystem: binaries in /sbin and /usr/sbin being > >demolished, random file corruption and whole directories being > >transplanted into lost+found. > I've never seen corrupted system files in the past 5 years - though > I've had reboots - usually because of my stupidity or carelessness. It must depend greatly on the hardware. I've got a system (4.7-RELEASE) that started having spontaneous reboots (with nothing at all logged to show why) about every night or two. I eliminated everything except the motherboard and CPU, and, interestingly, it stopped doing it when I stopped its nightly cvsup of the source and ports. Apparently that stressed it just a little too much. Anyway, when the machine would come back up, it would always complain of disk corruption and require me to manually fsck it. The problem was almost always with the / filesystem, even though /tmp, /var, and /usr are all much larger. (But they have softupdates set.) And the corrupted files would be files that definitely were not being written at the time of the crash like startup files in /etc, and one time it trashed /sbin/fsck itself. So there was something going on with that system or its drive (IDE) that caused corruption on a crash, even to parts of the disk that it had no reason to be writing to at the time. I've never seen anything like it with any other FreeBSD (or Linux) machine, or even with that machine until recently. -- Aaron From nobody Thu Feb 20 06:44:40 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FFS rant Organization: ESC Date: 20 Feb 2003 06:44:40 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 "Ivan Voras" writes: > Weird! Just the other day, a sysadmin friend of mine told me about > exactly the same behaviour on one of his boxes: critical > failure/reboot with nothing logged, in the middle of cvsup, very > repeatable. Also the fs areas that got somewhat trashed should not > have been written to. He suspected a HD failure (the box and the > drives are somewhat old) and even inquired on the stable@ > maillist. Maybe it's "just" a bug in the -stable? Mine is running 4.7-RELEASE. What seems odd is that I've got a modem connection, so I wouldn't think cvsup could bring in data fast enough to stress anything. I can buildkernel and buildworld on this box without any problems, and I'd expect them to be far more stressful. Even leafnode, the news server I run on it, should hit the filesystem harder when expiring than cvsup, I would think. I do assume it's ultimately a hardware problem, though. The drive isn't very old, but the MB/CPU are. -- Aaron From nobody Thu Feb 20 13:10:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: My FreeBSD 4.7 box keeps crashing..... References: Organization: ESC Date: 20 Feb 2003 13:10:11 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 36 "_Jung" writes: > What could be the problem? Is there any way for me to find out what > happened just before the crash? After a crash since the computer > locks up displaying a monitor-full of white vertical strips, I have > no idea if whether displayed some error messages just before the > crash.... Start a script like this running: while :; do w echo df echo vmstat echo netstat -m echo dmesg | tail echo ps -auxww echo sleep 30 done >logstuff That'll log a bunch of stuff to the file "logstuff" every 30 seconds. (Make sure you put "logstuff" somewhere where there's plenty of room.) After a crash, you can check the end of that file to see what was running, the load average, memory and disk usage, network memory buffers, kernel messages, and so on. -- Aaron From nobody Thu Feb 20 16:28:48 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: My FreeBSD 4.7 box keeps crashing..... Organization: ESC Date: 20 Feb 2003 16:28:48 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 kargl@troutmask.apl.washington.edu (Steven G. Kargl) writes: > Hmmm. Wouldn't be easier to add "options DDB" to the kernel > configuration file, build and install a new kernel, and then > when the system crashes he can do "ps" and "trace" at ddb> prompt. > With the "ps" and "trace" info, the developers may actually > be able to find and fix the problem. Probably. I do much of my work on remote systems, though, so I usually try not to tinker with the kernel unless I really need to, and in those cases I usually can't use a solution that doesn't allow the system to reboot automatically. I suspect that he'll either see disk or memory errors in the logs or dmesg output right before the crash, or he'll see nothing, in which case it's probably CPU or motherboard. -- Aaron From nobody Thu Feb 20 16:33:27 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Blog software References: <851c45c4.0302201146.478814b4@posting.google.com> Organization: ESC Date: 20 Feb 2003 16:33:27 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 brk_lists@yahoo.com (brk) writes: > I am looking for a Blog (weblog) software to install on a freeBSD box > running Apache + PHP + mySQL. Looking for something which is > multi-user and allows posting by e-mail. I believe PHP-Nuke fits your requirements, except possibly for the posting by e-mail. I've installed it, and it looks pretty good, but I haven't really used it yet. It's in the ports tree. -- Aaron From nobody Fri Feb 21 06:31:43 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: possible to scan a file and remove specific characters? References: <3e559897_1@aeinews.> <3E55B19A.5040805@tenebras.com> Organization: ESC Date: 21 Feb 2003 06:31:43 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 Michael Sierchio writes: > Usually all you need to do is cat the file: % cat a > b. The > stdin/stdout line discipline will remove the extree chars. It'd better not! How's cat supposed to know whether those characters are 'extree'? Cat knows nothing about text; it just streams bytes. > Otherwise you can use 'tr' or 'sed' or any number of tools. Or perl: perl -pi -e 's/\r\n$/\n/' file1 file2 file3 -- Aaron From nobody Sun Feb 23 15:49:07 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: GNOME or KDE References: <20030222211021.19428.00000053@mb-mj.aol.com> Organization: ESC Date: 23 Feb 2003 15:49:05 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 28 Stephen Paskaluk writes: > I never claimed they were. I have used and like KDE, and you'll > note that I am always eager to help people with their KDE questions. > But, and this is just IMHO of course, KDE is just too much stuff I > don't use, so I use Fluxbox. This still might be confusing the newbies, because you're comparing a window manager (Fluxbox) to a desktop environment (KDE), which also happens to include a window manager (kwm?). I don't think it's always apparent to newbies that they can use a lightweight window manager while still taking advantage of all the KDE and Gnome apps out there. I've been running WindowMaker for my window manager for a while. I also run Galeon, a Gnome-based web browser. A couple of the games I've installed are KDE-based. Some applications appear to be built on older desktop environments, like Motif or Athena widgets. I could also run all these apps in any other window manager. So if I've got a point here, it's that folks with slower machines shouldn't be afraid to install the big DEs like Gnome and KDE. If you find that their window managers, with all their extra features, are too much for your machine, you can certainly switch to a lighter one. -- Aaron From nobody Mon Feb 24 06:55:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: GNOME or KDE Organization: ESC Date: 24 Feb 2003 06:55:39 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 53 johntromaville@aol.com (JohnTromaville) writes: > To be honest I much prefer the window manager style of things than I > do the KDE (i'm picking on KDE cos I used it on a friends linux - > have never used Gnome). Still got a lot of playing to do but I > really do prefer window managers. You're always using a window manager when you use X. The Gnome and KDE packages include window managers, which is where it can get a little confusing. > You say that you can still take advantage of KDE and Gnome apps - > how easy is this. I mean do I have to jump through hoops to get them > to work or is it straight forward? I'm thinking mainly of KDevelop > but also any of the other apps. You just run them. They've been compiled against that desktop environment's libraries, which will be used within that application. In the FreeBSD ports tree, if you install something that needs Gnome or KDE, and you don't already have that DE installed, it'll be installed for you. But that doesn't change your window manager. There's a non-obvious separation between what WMs and DEs do. Put simply, a window manager is a program that manages the windows on your desktop. That means it controls where they go, how they layer on top of each other, how you move and control them, and what borders they might have. Some window managers also add extra features like docks, toolbars, menus, and other things. But everything your WM does happens 'outside' your applications. If I switch from windowmaker to fvwm, my apps themselves will all look exactly the same, but the borders and title bars may change, and the methods for iconifying and moving them around may change. A DE, on the other hand, is a set of libraries and an API for building graphical applications. Let's say you want to design a web browser. You could design all your own graphics and menu functions and scrollbars and so on, or you can use a DE that already provides those things ready-made. That allows you to say, "Create a File menu, with New, Open, and Quit options, which call these routines. Stick a scrollbar on the right side." The DE libraries provide a defined way to do those things. So if you design your browser using Gnome, it'll have a different look than if you use KDE, or Athena widgets, of LessTif, or some other DE. Newer DEs like Gnome and KDE go further than just offering graphical routines, providing sound routines and other libraries. But all these things effect what happens within the application, or 'inside' the window, if you will. -- Aaron From nobody Mon Feb 24 07:04:28 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What should I do with extra PC's? References: Organization: ESC Date: 24 Feb 2003 07:04:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 29 "_Jung" writes: > The problem is... I don't have requirements. I guess what I am > saying here is that I don't know what I need. Actually, I am even > sure if I have any "need" at all.... You could always just keep them around for parts, then. It's really impossible to say what you might need them for. I use an extra old machine as a headless backup server; my home directory gets rsync'd to it every hour. When my main system was very slow, I used to run mozilla from a second headless system, displayed over X. If you run a database, that would be a good app for a separate system. > Then you might say, why bother trying to use them while there really > isn't any need? Well, I don't know... When I see these old > computers taking up space in my basement doing nothing, I just feel > some sort of obligation to make some use of them. I know. It's > rather silly.... I've still got HP/Apollo 68040 33Mhz systems that I can't quite part with. I keep thinking there has to be some use for those huge monitors, at least. -- Aaron From nobody Mon Feb 24 07:14:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Configuring 2 NIC??(FreeBSD 4.7) References: <3e566751_5@news.bluewin.ch> Organization: ESC Date: 24 Feb 2003 07:14:04 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 "kaykay" writes: > Does that mean that I can't use two NICs if the two cards use the > same subnet?? It doesn't make sense. If the cards are in the same subnet, how does the system know which direction to send stuff to that net? The whole point of network addresses is to say, "This stuff goes here and that stuff goes there." If you have the same subnet on both sides of your machine, it should really be a bridge, in which case the NICs should have the same IP, not just the same network address. -- Aaron From nobody Thu Feb 27 06:27:15 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What should I do with extra PC's? References: Organization: ESC Date: 27 Feb 2003 06:27:13 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 "_Jung" writes: > I have two computers as of now. One has P200, 98MB RAM, 3.2 GB hard > drive, and 100/10 PCI NIC, and the other has P133, 32MB RAM, 6.2 GB > hard drive, and slow (10) ISA NIC. > > The parts are, of course, interchangable. I just set these two > computers up this way because I thought a firewall should have a > faster CPU and a faster NIC with more meory, while a file server > should have a bigger hard drive and can get away with a slower CPU > and a slower NIC. Am I right on this?? I'd say just the opposite, although it depends on your situation. There's no need for your workstation-firewall connection to be faster than your firewall-Internet connection, so unless you've got a huge pipe to the Internet, your 10 Mbit card will never be loaded. On the other hand, if you transfer many large files to/from your NFS server, the faster card will make a difference there. The choice of CPU and RAM probably won't make as big a difference, but I'd put the better ones in the NFS server. That's simply because if you add other services to a machine in the future (as you mention later), you probably won't put them on the firewall. And firewalling is not CPU- and RAM-intensive in the least. > Since it's going to be a network of one internal user, I thought a > 10MB NIC would be good enough connection for file sharing > (NFS/Samba) and NIS. It's enough (it's what I'm using) but 100 would be better. -- Aaron From nobody Thu Feb 27 06:37:28 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: X Window seesion in a window References: Organization: ESC Date: 27 Feb 2003 06:37:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 32 Dave Goode writes: > I'd like to be able to have an X session to a remote computer > running in a window in my existing local X session. Is there > anything around that can do this? Running X clients displayed on your local display is built into X. You have to do two things to make it work: 1. Your server (local display) must allow it. This can be done a few different ways, from xhost to various cookie and authentication methods. To try out the simplest, just do this: xhost +remote_machine Once that works, you can look into more secure methods. 2. Tell the remote client what server to connect to. Just about every X client understands the -display option, like: xterm -display my_server:0.0 That runs an xterm on whatever machine you execute it on, displaying it on the first display on my_server. You can make much of this transparent by using X11 tunneling in ssh; see the ssh and sshd man pages about that. -- Aaron From nobody Fri Feb 28 09:11:34 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What a typical FreeBSD network looks like? References: <86r89s4lqi.fsf@vanilla.zzz> Organization: ESC Date: 28 Feb 2003 09:11:34 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 32 Peter Wu writes: > My company runs almost everything on Microsoft products. There are > DC's, Windows workstations, Active Directory, Exchange, etc.. > I'd like to know what a typical FreeBSD network environment looks > like. I mean everything is built on FreeBSD. How are the networked > FreeBSD boxes connected to each other? Do they also live in a > Domain? Is there a Single-Logon feature? All the same concepts are there (and have been for decades), but with different terminology. Users and hostnames can be centrally controlled with a service like NIS. IP addresses can be centrally controlled and handed out by bootp. Home directories can be shared across machines with NFS. So yes, in a well-configured Unix network, you can sit down at any machine (assuming you're allowed to), login, and have your own environment and home directory. Thanks to X, you can also run applications on any number of other machines and have them displayed where you're sitting. The biggest difference I've seen between a Windows network and a Unix one is that the Unix network doesn't need as many server machines, because the services aren't nearly as bloated. Also, in my experience, the Unix network needs far fewer administrators. (Unfortunately, for us Unix admins. :-)) -- Aaron From nobody Fri Feb 28 21:31:17 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: What a typical FreeBSD network looks like? Organization: ESC Date: 28 Feb 2003 21:31:17 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 Peter Wu writes: > > So yes, in a well-configured Unix network, you can sit down at any > > machine (assuming you're allowed to), login, and have your own > > environment and home directory. Thanks to X, you can also run > > applications on any number of other machines and have them > > displayed where you're sitting. > Say, I log on machine A with my username/password and want to access > a share on machine B. Do I have to type my username/password again? Not unless you go out of your way to make it that way. Normally, whatever NFS mounts you have configured would either be mounted all the time, or set up to automount when you access them. It works at the filesystem level, rather than as a user process like SMB typically does. -- Aaron From nobody Sat Mar 1 05:46:46 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: starting dual named daemons References: Organization: ESC Date: 01 Mar 2003 05:46:46 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 15 David Vidal Rodríguez writes: > Maybe you can address the problem by copying the script > /etc/rc.d/named to /etc/rc.d/named2 and changing the variables in > there (name=named2, $named_chrootdir -> $named_chrootdir2, etc.), > and then adding a line in /etc/rc.conf Um, there's no /etc/rc.d in FreeBSD. Bind is started from rc.network. If I wanted to start a second instance of named, I'd write a shell script to do it, and put it in /usr/local/etc/rc.d. -- Aaron From nobody Sat Mar 1 07:56:34 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: starting dual named daemons Organization: ESC Date: 01 Mar 2003 07:56:34 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 18 David Vidal Rodríguez writes: > $ ls -ld /etc/rc.d > drwxr-xr-x 2 root wheel 2,0K 21 ene 19:07 /etc/rc.d/ > > $ uname -v > FreeBSD 5.0-RELEASE-p1 #6: Tue Feb 25 20:45:45 CET 2003 > > Maybe it's different in 5, but i *do* have an rc.d. Ah, that's the difference. I'm using 4.7. Seems a little SYSVish to me. :-) -- Aaron From nobody Sun Mar 2 06:16:52 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: starting dual named daemons Organization: ESC Date: 02 Mar 2003 06:16:52 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 "cyber_cynic" writes: > I can start the daemon from the command line with : > named -t /chroot/bindint -c named.conf -u bindint > So I took the script is for msql client (000.mysql-client.sh) and > adapted it as follows : It looks fine to me. You may want to use a full path to named, just in case it isn't in the path that's in effect on startup. > Will that work if I called it say 000.bind-daemons.sh It'll be run (with the argument "start") on boot if it's in /usr/local/etc/rc.d, ends in ".sh", and is executable. (I'm not even sure if that third part is necessary.) -- Aaron From nobody Sun Mar 2 17:41:16 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Is there a 'Production' branch in FreeBSD? References: <5N68a.9485$x_5.1068987@news20.bellglobal.com> <2Mr8a.3499$P65.47811@weber.videotron.net> Organization: ESC Date: 02 Mar 2003 17:41:16 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 Mina Naguib writes: > I think it would work a lot more cleaner and elliminate a lot of > beginner's confusion if, when building from -CURRENT or -STABLE, they > identify themselves with treename-synctime such as "FreeBSD > STABLE-200302280400" which exactly means "I'm running the STABLE tree > as it was at this point in time". Verion numbers such as 4.7 only > belong and make sense in releases. Yes. After all, box A running 4.7-STABLE and box B running 4.6-STABLE may be more similar than box A is to box C running 4.7-STABLE from a different cvsup date. -- Aaron From nobody Mon Mar 3 06:37:28 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: First week with FreeBSD References: <4iq8a.30653$em1.7947@news04.bloor.is.net.cable.rogers.com> Organization: ESC Date: 03 Mar 2003 06:37:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "Keith Boone" writes: > 2) Browser Plugins. I admit I haven't even had time to try > installing plugins yet, but I've read that a lot of people seem to > have problems. I'd just like one that'd get my browser (Galeon) to ignore Flash. I tried using the Flash plugin, but it crashed the browser too often. So now I have to click 'cancel' on a popup every time I go to a Flash page, even if I'm just backing up to it. -- Aaron From nobody Mon Mar 3 06:44:15 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: nvidia geforce2 chipset with xfree86 4.3 References: Organization: ESC Date: 03 Mar 2003 06:44:15 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 "Jose Yimpho" writes: > I'm running 4.7 with Nvidia's proprietary binary drivers for my > nforce geforce2 board. Apparently, XFree86 4.3 has support for the > integrated nvidia nforce2 chipset. > Will XFree86's driver be better than Nvidia's? XFree86's xv driver didn't work with my integrated NVidia motherboard (Asus A7N233-VM) -- it locked the machine up tight every time I started it. The nvidia driver works fine. -- Aaron From nobody Mon Mar 3 06:47:31 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: mail.cgi References: Organization: ESC Date: 03 Mar 2003 06:47:31 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 "BSD" writes: > I am running qmail on FreeBSD ... it works fine > (user@bsd)(~) />mail: /usr/sbin/sendmail: No such file or directory > how can i make it to use qmail as mail transfer agent. You should follow the INSTALL doc that comes with qmail (or even better, follow the instructions at lifewithqmail.org; that sets it all up with daemontools and nice logging, and supplies you with some good scripts). Then: ln -s /var/qmail/bin/sendmail /usr/sbin/sendmail -- Aaron From nobody Mon Mar 3 15:03:26 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: nvidia geforce2 chipset with xfree86 4.3 Organization: ESC Date: 03 Mar 2003 15:03:26 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "Jose Yimpho" writes: > > XFree86's xv driver didn't work with my integrated NVidia > > motherboard (Asus A7N233-VM) -- it locked the machine up tight > > every time I started it. The nvidia driver works fine. > Do you mean XFree86's nv driver? Yes, that's what I meant. -- Aaron From nobody Tue Mar 4 15:52:52 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /usr file system full References: Organization: ESC Date: 04 Mar 2003 15:52:50 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 59 "Ernest Hammerschmidt" writes: > I'm having a strange problem and I just can't seem to fix it. I was > looking through man pages (logged in as a "normal" user, no root > privileges) and by mistake did a > > man mount |< less > which very soon greeted me with fsclose: file system full messages. > 2) Why do I seem to have filled /usr, though I shouldn't even have > write access? Are you sure your home directory isn't on /usr? I think they typically are by default on FreeBSD. If that's not the case, there might be tmp directories that you can write to. /tmp is usually that way, so you could have /usr/tmp or something like that setup that way. Since you were running 'man mount' at the time, check /usr/share/man/cat8 and similar places. The man command saves a formatted copy of each man page in a 'cat?' directory the first time it reads it, so subsequent reads will be much faster. If you somehow made man go haywire, it could have saved a huge file there. I kinda doubt that, though, since man runs as the user 'man', which shouldn't be able to overfill the disk like that. In fact, only root should be able to do that. > 3) What can I do about it? Well, since your filesystem is 500MB+ over capacity, you've probably got a file at least that large somewhere. Here's a perl script that'll find it: --------------------------------------------- #!/usr/bin/perl -w use strict; use File::Find; find( \&wanted, "/usr"); sub wanted { my $size = -s $File::Find::name; if( $size > 500000000 ){ print "$File::Find::name - $size\n"; } } --------------------------------------------- (I figured the 'find' system command would have a 'bigger than' test, but it doesn't, so this script will do the job.) -- Aaron From nobody Tue Mar 4 16:04:12 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /usr file system full References: Organization: ESC Date: 04 Mar 2003 16:04:12 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 "Ernest Hammerschmidt" writes: > Because when I scroll down using the arrow key, man exits when I > reach the end of the man page. Using less was just a try to prevent > this from happening. Set the PAGER variable in your environment to "less". In sh, bash, ksh: PAGER=less; export PAGER In csh, tcsh,: setenv PAGER less -- Aaron From nobody Tue Mar 4 16:08:46 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /usr file system full References: <20030304213636.17136d52.steveo@eircom.net> Organization: ESC Date: 04 Mar 2003 16:08:46 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "Ernest Hammerschmidt" writes: > > EH> man mount |< less > Absolutely. Just checked my .bash_history. That's what it is. What shell were you using? I just tried that in sh and csh, and neither one would 'work.' Csh said "Ambiguous input redirect" and quit right away. Sh said "cannot open less: no such file", and then man complained because it couldn't output to anything. -- Aaron From nobody Tue Mar 4 20:03:05 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /usr file system full References: Organization: ESC Date: 04 Mar 2003 20:03:05 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 25 "Ernest Hammerschmidt" writes: > Well, besides the usual stuff, WindowMaker and grip. But that should > be about it. Of course, it _might_ be that the problem existed even > before I did that strange pipe thing. So this might have just > uncovered the whole thing. If I rerun the command, I'm getting the > same effect, but the amount of used disk space doesn't grow. I think you may be right. 'man mount' tries to write a pre-formatted file to /usr/share/man/cat8, so that would trigger the filesystem full message. Something else may have already filled it. You could use 'find' to list all files created or modified within a certain time period. For example: find / -ctime -2h # finds everything changed in the last two hours find / -ctime -1d # finds everything changed in the last day Since you didn't find one huge file, this might help you find a big bunch of small ones. -- Aaron From nobody Tue Mar 4 20:08:05 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: X Issue from Remote Client References: <3e6545f3@news.isdn.net> Organization: ESC Date: 04 Mar 2003 20:08:05 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 Walt Harris writes: > In the office, I have a FreeBSD 4.7 box and a Slackware Linux 8.1 > box both running Apache, PHP, PostgreSQL, and SSH. If I connect to > the FreeBSD box from Slackware, and 'xterm -display SLACKWARE:0', I > get a xterm window up and I can proceed to do work. If I connect to > the Slackware box from FreeBSD, and 'xterm -display FREEBSD:0', I > get a "can't open message..." as if it couldn't find the name of the > system in the access control list. On your FreeBSD box, do 'xhost +SLACKWARE'. That'll let clients running on SLACKWARE connect to your server. Or you could use X11 Forwarding in your sshd configs, but I don't bother with that on my local network, since it seems wasteful to encrypt all that traffic where there's no security issue. -- Aaron From nobody Wed Mar 5 07:08:29 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /usr file system full Organization: ESC Date: 05 Mar 2003 07:08:29 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Erik Nygren writes: > > (I figured the 'find' system command would have a 'bigger than' > > test, but it doesn't, so this script will do the job.) > Oh, but it has. > This one looks for files larger than 10k: > find . -size +10240c Ah, thanks. I thought that was odd that I couldn't find that. I saw the +/- options to the time tests, but not for size. -- Aaron From nobody Wed Mar 5 07:11:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Perl References: <2603221.1046815575@dbforums.com> <20030304192247.G97179@localhost.my.domain> Organization: ESC Date: 05 Mar 2003 07:11:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 casaveli@hotmail.com writes: > On Wed, 5 Mar 2003, FreeBSD/Ghosted wrote: > > Jose Yimpho wrote: > > >> pons wrote: > > >>> I am currently running freebsd 4.4 stable, Perl was installed > > >>> by default > > > I am currently running freebsd 4.7 stable. The water outside is > > > wet. > > I am currently running windows. My hair is missing. > I am currently running FreeBSD 4.8-RC. The sky is blue. "If it weren't for my horse, I wouldn't have spent that year in college." -- Lewis Black -- Aaron From nobody Wed Mar 5 20:37:37 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: freebsd and ntfs References: <44u1ehpihq.fsf@be-well.ilk.org> Organization: ESC Date: 05 Mar 2003 20:37:37 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 mike writes: > all i want to do is replace my mozilla directory in freebsd with a > symlink to my mozilla directory in windows. that way, i can get my > email from both sides. I think the usual way to do this is to create a VFAT or DOS partition, and keep all your shared stuff there. Windows and FreeBSD can both use those formats. Of course, that assumes that the two versions of Mozilla will happily share files. I don't know about that. -- Aaron From nobody Thu Mar 6 06:29:17 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: freebsd and ntfs Organization: ESC Date: 06 Mar 2003 06:29:16 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 mike writes: > so just out of curiousity, is there any particular reason why > freebsd's ntfs driver doesnt have complete write capabilities? Because you haven't written it yet. :-) With PCs being as cheap as they are these days (especially ones powerful enough to run FreeBSD), I would suspect that dual-booting is becoming less common. With a second PC, a couple network cards, and Samba, you can keep both OSs running, which is much more convenient than dual-booting. That also lets you do some other handy things, like keeping your Windows machine behind your firewall, and backing the two systems up to each other. -- Aaron From nobody Thu Mar 6 14:22:51 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: /etc/login.conf References: <3E67A103.8060304@piaski.eu.org> Organization: ESC Date: 06 Mar 2003 14:22:51 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 28 highland writes: > I edited in default section PATH, then reloaded databases > but still have old $PATH (even after reboot), why ? Are you running csh or tcsh for your shell? Those are the shells that use login.conf. You might post what you added so we can check it out. > I tried to add: > PATH=$PATH:/new/path: Don't know if it'll hurt anything, but you don't need that trailing colon. > export $PATH Should be "export PATH". When you include the $, the variable is replaced by its value, so it tries to export a variable called "/bin:/usr/bin:/new/path" or whatever the value of PATH is. > to /etc/profiles Should be /etc/profile. -- Aaron From nobody Fri Mar 7 05:21:38 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: (possibly) a dumb question about groups and unix References: Organization: ESC Date: 07 Mar 2003 05:21:38 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 32 David King writes: > Users have a primary group, and then several secondary groups. For > instance, a user may be a member of the group "users," as their > primary group. But they also are members of the groups "wheel," > groups that give them access to certain nfs mounts, or whatever > else. When they create a file, it is owned by their user, their > _primary_ group, Not quite. The group on a new file is set to the same group as the directory it's in, if it's one of your groups. If it's not, then I assume it defaults to your primary group. For example, I'm in two groups, abaugher(primary) and wheel. So: $ mkdir abaugherdir; chgrp abaugher abaugherdir; touch abaugherdir/test $ mkdir wheeldir; chgrp wheel wheeldir; touch wheeldir/test $ ls -l */test -rw-r--r-- 1 abaugher abaugher 0 Mar 7 05:16 abaugherdir/test -rw-r--r-- 1 abaugher wheel 0 Mar 7 05:16 wheeldir/test That's pretty handy, since it means that if several people share a group, you can create a directory for them to share their work in, chgrp it to the shared group, and everything that they create under that directory from then on will have that group ownership, without them having to manually set their group. -- Aaron From nobody Fri Mar 7 08:12:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: qmail problem References: <3E689647.7040608@piaski.eu.org> Organization: ESC Date: 07 Mar 2003 08:12:10 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 highland writes: > Hello > I am installing qmail from ports: > and i see 5 qmail proceses, but > nobody listen on any interface on port 25, > firewall disabled, no errors in logs. Read the part in the qmail FAQ or INSTALL file about setting up qmail-smtpd with tcpserver. Or better, follow the instructions at lifewithqmail.org. -- Aaron From nobody Sat Mar 8 15:15:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: making world and kernel References: Organization: ESC Date: 08 Mar 2003 15:15:04 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 10 "haloflightleader" writes: > I'm thinking of installing it on my other laptop. Thoughts? I like pie. -- Aaron From nobody Sat Mar 8 15:25:00 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: portupgrade - failed portsdb -Uu References: <87vfytoopd.fsf@strauser.com> Organization: ESC Date: 08 Mar 2003 15:25:00 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 "cyber_cynic" writes: > > Try: > > > % portsdb -Uu & > Thanks for the tip, I will try it but would like to add something if > possible - is there a way to issue the command so anything that > would have been written to the screen (assuming it was not running > in background mode) is written to a file. Sure: sh -c "portsdb -Uu >~/portsdb.log 2>~/portsdb.err" & That will redirect standard output to portsdb.log, and errors to portsdb.err, both in your home directory. -- Aaron From nobody Sat Mar 8 15:32:05 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Backup to hd: dump vs. dd References: <7jic4b.77.ln@luva.home> Organization: ESC Date: 08 Mar 2003 15:32:05 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 migieger@web.de (Michael Giegerich) writes: > I'm backing up (better copying) my FreeBSD-4.7 system > from the scsi disk it resides on to an ide disk in the > same PC (yes, I'm aware that it isn't a real backup. > To do the real thing I use a Tandberg SLR tape backup). > Using dd has the huge advantage that it's *fast* and > the disadvantage that it mangles the disklabel entry > of the target disk. > Dump otoh leaves the disklabel as it is, but is dead > slow (hours vs. minutes or ~1 MB/s vs. ~25 MB/s). > As I would like to use dump, the obvious question is > how to speed up the dump process. Any hints? (Alterna- > tively hints for using dd while keeping the target > disklabels intact would be welcomed to). Dump is slower because it checks every file and only backs up the ones that have changed since the last dump of the same level. So if you're always doing a full backup (level 0), there's no point in using dump; you might as well just dd the whole thing. I'd be surprised if a dd from a SCSI disk to an IDE disk creates anything usable, but I guess it's possible. In cases where I'm backing up to a second disk or another machine on the localnet (not something removable like a tape, in other words), I'll setup an rsync command to run occasionally. That way it only backs up what's changed, and it can also keep the second disk a very close copy of the first one, so that if the first disk dies, I can boot from the second one very quickly. That's not a substitute for a removable, off-site backup, but it's handy for when you just realize you trashed some files and need to go get them quickly, since they're right there on the other disk/system, rather than tied up in some archive. -- Aaron From nobody Mon Mar 10 14:21:19 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Use the -display ip command References: <3E6C6D2E.4070405@planet.nl> Organization: ESC Date: 10 Mar 2003 14:21:18 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 J Vendrik writes: > I have seen that people use the -display command. > xeyes - display > So you can run the program at the local pc and show it to the > What must i configurate and how works it ? Let's say you want to run xeyes on a machine whose name or IP address is REMOTE, and display it on a machine called LOCAL. On LOCAL, once X is started, you need to run this command: xhost +REMOTE If you expect to do this often, put that command in your .xinitrc or .xsession on that machine. Then on REMOTE, you run this: xeyes -display LOCAL:0.0 -- Aaron From nobody Mon Mar 10 17:52:38 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: irc References: Organization: ESC Date: 10 Mar 2003 17:52:38 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 "Gunnar" writes: > where can i find a howot setup a irc server, and whats the ports > name ? How to do it is going to depend on which server you use, and which network you want to connect to. There appear to be a few ircd ports in /usr/ports/irc. You could choose one, install it, and then see what its documentation says about getting connected to a network. Or, if you already have a network selected, see if its web site has any suggestions on what server to use, or connect to that network and ask in a help channel. -- Aaron From nobody Thu Mar 13 05:59:28 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: SU not enabled References: Organization: ESC Date: 13 Mar 2003 05:59:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 Huub writes: > Ok, thank you...and how do I insert the user into the wheel group? # pw user mod username -G wheel If the user needs to be in any secondary groups other than 'wheel,' add them after 'wheel' separated by commas and no spaces: # pw user mod modemuser -G wheel,dialer The 'pw' command is safer than editing the passwd/group files directly. -- Aaron From nobody Fri Mar 14 06:46:33 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: cannot install XEmacs with GTK References: Organization: ESC Date: 14 Mar 2003 06:46:33 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Peter Wu writes: > I attempted to install XEmacs with GTK from ports but no luck. :( > > # cd /usr/ports/editors/xemacs-devel-mule > # make -DWANT_GTK I don't know if this relates to your problem, but that's not how to define something when installing a port. Do it like this: make WANT_GTK=yes -- Aaron From nobody Mon Mar 17 18:51:08 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: BSD Project - something to sink your teeth into? References: <8W7da.682621$Wr.25155964@Flipper> Organization: ESC Date: 17 Mar 2003 18:51:07 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 60 Daemon writes: > X is confusing this way... I want the server to be able to supply > the windows manager on my laptop/desktop whatever with an X > service. Pretty sure that what I wrote is correct. I do realise > that, unlike most client/server applications X requires the > 'powerful' part to be run on the client (the one with a good > graphics card etc) and the server part can run on just about > anything... I think. if not then I am going to have to do some > serious reading on the XFree site. You've got the right idea, but you're still getting the client/server terminology wrong. Don't feel bad; nearly everyone does at first, because it feels backwards. An X server is a program that runs on a machine and controls the display and inputs of that machine. This will almost always be the machine you're sitting in front of. The machine itself may be a full-blown Unix system running X from disk, or a proprietary diskless X-Terminal running X from ROM, or a Windows PC running an X server like X-Win32, or some other setup. An X client is a program that talks to an X server. It usually asks the X server to display something, but even that's not always true: xwininfo requests information from the server without displaying anything. But usually an X client asks the X server to display something and pass along inputs from the keyboard and mouse that apply to that display. X clients may run on the same physical machine as the server, or on other machines. This networking is one of the most important features of X -- clients always talk to the server through network sockets over some sort of networking protocol. You'll want to put your best graphics card on the server machine, along with your best monitor, keyboard, and mouse. Where you put your other most powerful components depends on where you plan to run your bulkiest clients. Usually you'll want your best CPU, RAM, and disks on the server too, but not necessarily. For example, lets say my server (the one I sit at) is called Alpha, and I have three other machines named Beta, Gamma, and Delta. I put my best graphics card in Alpha, along with enough RAM that it'll swap as little as possible. I plan to use Beta for all my compiling, so I put my fastest disks in that machine. The other two machines will be for CPU-intensive apps like web browsers and games, so I'll make sure they have plenty of CPU power. You have a choice of a few authentication systems for controlling how your X clients are allowed to talk to your server. You can use the X11 Forwarding built into ssh, but on a local network, the overhead of encrypting all that traffic is probably an unnecessary waste. There are other methods involving cookies, but the simplest way is to use the xhost command on the server to specify which client machines are allowed to connect. -- Aaron From nobody Wed Mar 19 14:00:37 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie Xwindows help References: <1rSda.24107$qL4.514833@twister.rdc-kc.rr.com> <20030319081248.0bf31716.steveo@eircom.net> <_73ea.25568$qL4.543345@twister.rdc-kc.rr.com> Organization: ESC Date: 19 Mar 2003 14:00:36 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 iso50 writes: > Thanks for the info on getting it going. The only problem is I > can't su while in Xwindows. Are you a member of group 'wheel'? Run the 'groups' command in an xterm and see what it says. If you're not, you'll need to login as root at a console and do: pw user mod USERNAME -G wheel Replace USERNAME with your username, and that'll add you to the wheel group. That should let you su. -- Aaron From nobody Thu Mar 20 07:33:29 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: creating PDF files References: Organization: ESC Date: 20 Mar 2003 07:33:29 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 Keve Nagy writes: > Does anybody know if there are tools or applications to create PDF > files under FreeBSD? The PDFlib library has APIs for several different languages. See /usr/ports/print/pdflib. > If yes, any link to a documentation on how to set these up? And > again if yes, how is this PDF different from the Adobe Acrobat PDFs. They should be compatible, but YMMV. Recently I sent a PDF file created with PDFlib to a Mac user. It was created with host fonts from my FreeBSD system. She was able to view it in Acrobat Reader, but it wouldn't print. If you stick with the default PDF fonts, that should help. -- Aaron From nobody Thu Mar 20 19:28:27 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: One computer as multiple X clients. References: <20030320085632.277e3775.steveo@eircom.net> Organization: ESC Date: 20 Mar 2003 19:28:27 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 Steve O'Hara-Smith writes: > Why should it not be possible ? With the right set of Device, > Monitor and InputDevice sections in XF86Config it should just work > AFAICT. That's what I would think, but I don't have a setup to try it on. The BusID option in the Device section of XF86Config should take care of the multiple video cards. The mouse section lets you set a device path, so that should be no trouble. I'm not sure about the keyboard, though; I don't see anything in the same XF86Config to let you point the keyboard to a specific piece of hardware, but there might be something in the man page. Ultimately, if the kernel recognizes all the hardware, I'd expect this to be doable. I'm not sure it'd be that useful, though. You're still going to have to have a monitor, keyboard, and mouse for each station. You'll probably need to add RAM to the system, and maybe have a faster CPU, to be able to keep up with running two X servers at once. It might be more trouble than it's worth just to avoid buying a second MB/case/PS. -- Aaron From nobody Fri Mar 28 14:26:39 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Adding a second IP to network interface References: Organization: ESC Date: 28 Mar 2003 14:26:38 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 "Maziar Aflatoun" writes: > I'm trying to add a second ip my network interface and it doesn't > work. Here is the part in my rc.conf > ifconfig_xl0_alias0="inet 10.0.10.45 netmask 0xffff0000" Set the netmask on your secondary addresses to 0xffffffff. -- Aaron From nobody Fri Mar 28 17:17:09 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: what does << do and why doesn't it work inside of an if References: Organization: ESC Date: 28 Mar 2003 17:17:09 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 48 Mike writes: > I'm trying to write a bash script. > I'm using: > ftp -n -i << END_FTP > open $mydomain > user $myusername $mypassword > cd $subdirremote > ls > mput * > bye > END_FTP > which works fine .. UNTIL I put it inside an if statement.. then I > get an error: line .. unexpected end of file. What would cause this? My guess is that you're indenting the END_FTP marker. That won't work; the shell expects to find it at the beginning of a line. So it looks all the way to the end of the file for it, gives up, and then can't find your closing 'fi' because it's past it already. If you really want to indent your ending marker, you can add the same spacing to the opening marker: if true; then cat << " END_FTP" Howdy.... END_FTP fi By using quotes, I added two spaces to the << marker, to match my closing marker later. > Maybe it would help if I understood what the << is really doing for > me. I thought it just told the script to use the following series > of commands until it ran across the specified string (END_FTP in > this case) but perhaps I misunderstood. Maybe a little. The << notation says to take the text between here (that's why it's called a 'here' document) and the matching closing text, and feed it to the standard input of the command. So your example feeds those lines to the ftp process that's started. It should work, once you fix the indentation. -- Aaron From nobody Sun Mar 30 07:07:47 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: time config References: Organization: ESC Date: 30 Mar 2003 07:07:47 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Matt writes: > i was wondering, where i could configure the timezone, daylight > saving start/stop date, GMT status. STOP don't tell me > /stand/sysinstall please. where is the config file? the freebsd > handbook didn't tell me. it just pointed me to /stand/sysinstall but > there i can't set the daylight saving start/stop date... All of that stuff is included in the timezone definition file. If you don't like any of the predefined timezones, you can compile your own. See 'man zic' for info. -- Aaron From nobody Mon Mar 31 06:57:43 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: local DNS References: Organization: ESC Date: 31 Mar 2003 06:57:43 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 pi@complx.LF.net (Kurt Jaeger) writes: > Using /etc/hosts with 5 systems is from my point more difficult than > setting up an DNS server, because it will be messy to keep in sync. You might want to look at YP/NIS. It'll let you centralize your host/IP configurations, and it'll do the same for your users and groups if you want. -- Aaron From nobody Mon Mar 31 20:03:30 2003 Newsgroups: comp.unix.bsd.freebsd,comp.unix.bsd.freebsd.misc Subject: Re: $ in pass for su References: Organization: ESC Date: 31 Mar 2003 20:03:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 "ME" writes: > My root account on my box uses a $ sign as the last character of the > password. My primary user account (say bob) does not. Other than > that the two passwords are identical. So say the pass for bob is > "big7monkey" and for root it's say "big7monkey$" (its not really but > for the sake of example lets pretend). Now the question is this: > When logging on via telnet, ssh, or even the console my main user > account (bob) can type the password (big7monkey) and just for kicks > can throw on the $ sign at the end and oddly the user is allowed > login. It sounds like your password encryption is only recognizing the first 8 characters. (Try using 'big7monk'; it'll probably work.) Your system is probably using DES encryption. You can change to MD5 encryption, which will let you use up to 128 character passwords by default, and longer if you change some settings. Check the handbook for more details. -- Aaron From nobody Tue Apr 1 15:38:02 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: need some help with ls command References: Organization: ESC Date: 01 Apr 2003 15:38:01 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 "Peter Pippinger" writes: > > % ls -lad filename.txt | awk '{print $5}' > sorry for asking again, but is there also a way to display only the > first result (if using something like ls file*) ? Pipe it through "head -1". -- Aaron From nobody Tue Apr 1 18:01:48 2003 Newsgroups: comp.unix.bsd.freebsd,comp.unix.bsd.freebsd.misc Subject: Re: Alternatives to NAT? References: <7Ooia.293380$L1.83657@sccrnsc02> Organization: ESC Date: 01 Apr 2003 18:01:47 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 "ME" writes: > Say you have so many users you must create a 10.0.x.x network. You > can not use NAT (NAT can't pass IPSEC which is a major requirement > for my office.), nor, from what I understand, can you use a bridge > (because of the 10.0.x.x network). What kind of "routing" would you > use? Eventually you would have to have some kind of router to route > traffic from the internet to the private network(s). Am I wrong on > this? What kind of approach would this entail and can FreeBSD > handle such a situation? That's the very definition of NAT. If you don't have enough public IP addresses for all the machines on your network, you use NAT. If you can't use NAT, you can't use a private IP range. Here's why: There's no way for machines in the outside world to route to your 10. machines. So they have to route to a public [A]ddress, or set of addresses, and then that/those have to be [T]ranslated to the 10. addresses upon arriving at your internal [N]etwork. -- Aaron From nobody Tue Apr 1 18:07:44 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: need some help with ls command Organization: ESC Date: 01 Apr 2003 18:07:44 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 Michael Sierchio writes: > Start another process and a pipe when you can do it all > in awk? Hiss! ;-> > ls -lad file* | awk 'FNR==1 {print $5}' I'd also use 'wc' instead of 'ls'. That way if the match is really a symbolic link, you get the size of the real file, instead of the size of the link. If that's what you want. wc -m file* | awk 'FNR==1 {print $1}' -- Aaron From nobody Tue Apr 1 19:07:53 2003 Newsgroups: comp.unix.bsd.freebsd,comp.unix.bsd.freebsd.misc Subject: Re: Alternatives to NAT? Organization: ESC Date: 01 Apr 2003 19:07:51 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 39 "ME" writes: > So then just about every Huge organization today uses NAT, in some > form or another either their Cisco routers or their software based > gateways? Yes, if they don't have enough public addresses for all the machines that need to access the outside. I've done work for companies that had full class B ranges, so they had plenty of addresses for everything and didn't have to bother with NAT. I've dealt with others that had plenty of addresses, but used NAT anyway because someone convinced them it increased security. > That must of been a finer point they did not teach me in the CCNA > class. I was truly under the impression that a router could simply > route from one network to another regardless of whether it was > private or not. That question has stumped me for some time. Thank > you. As far as the router itself is concerned, the question of public vs. private doesn't really matter. You can certainly setup any router to route from 10.0.0.0/8 on one side to 0.0.0.0 on the other side. The router itself will route between the two just fine. The problem is that responses and incoming connections will never get to the router, so that's where NAT comes in. A while back, I had four machines and no hub in my office, so I setup three private networks daisy-chained. An Internet-connected PC connected on 192.168.1/24 with crossover Ethernet to an Apollo Domain system, which connected on 192.168.3/24 with Domain Ring to another Apollo, which connected to another PC with crossover Ethernet on 192.168.5/24. With a few static routes, they all talked to each other just fine, but to talk to the outside, they all used NAT (except for the one with the public address). -- Aaron From nobody Wed Apr 2 04:43:24 2003 Newsgroups: comp.unix.bsd.freebsd.misc,alt.comp.periphs.mainboard.asus Subject: Re: [FreeBSD] Which LAN driver for ASUS A7N8X (nforce)? References: Organization: ESC Date: 02 Apr 2003 04:43:23 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 "Andreas Berg" writes: > As far as I can tell, the nVidia nForce MCP is not supported at this > time. That was my conclusion. I just use a PCI NIC with my Asus A7N266-VM, and I think I disabled the MCP in the bios, although that probably doesn't matter. -- Aaron From nobody Wed Apr 2 05:06:27 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: DJBDNS & Tinydns References: <2ock8vo2hi1ggcn26u63801uek57o846p4@4ax.com> Organization: ESC Date: 02 Apr 2003 05:06:25 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 50 North-- writes: > i cant make sense of these log files cached > @400000003e8a2b542b2c7b8c c0a80001:0fcd:01c2 - 000c > 86.105.168.216.in-addr.arpa The parts that I'm sure of are: @400000003e8a2b542b2f36c4 -- The 64-bit time of the loookup. You can pass this through tai64nlocal to convert it to localtime. c0a80001 -- IP address of the system doing the lookup, in hex notation. This one converts to 192.168.0.1. Here's a little perl script you can use to convert 8-character hex strings into dotted-decimal IPs: ---------------------------------------------------------- #!/usr/bin/perl -w use strict; die "Usage: hexip [ffffffff] ...\n" unless $ARGV[0]; for (@ARGV){ print "$_ "; if( /^([0-9a-fA-F]{2})([0-9a-fA-F]{2})([0-9a-fA-F]{2})([0-9a-fA-F]{2})$/ ){ print join '.', hex($1),hex($2),hex($3),hex($4); } else { print "Invalid"; } print "\n"; } ------------------------------------------------------------ :0fcd:01c2 - 000c -- Dunno about these parts. Probably source port, something else, and a status code of some sort. 86.105.168.216.in-addr.arpa -- The thing being looked up. This is a reverse lookup of an IP. > where are the ip addresses? and how do you figur them out? > it seems some caches are repeated, why? Tinydns doesn't act as a cache; that job is done by dnscache. So if tinydns gets several requests in a row for the same lookup, it'll log each of those requests normally. -- Aaron From nobody Wed Apr 2 10:18:31 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: DJBDNS & Tinydns Organization: ESC Date: 02 Apr 2003 10:18:28 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 42 Michael Sierchio writes: > Install daemontools and use > > # tai64nlocal < {logfile} Yes, as I said, tai64nlocal converts the 64-bit times to human-readable time. But it doesn't do anything to the hex-notation IP address, which is what that little script I included does. Here's another script that does both; just give it the logfile as an argument. --------------------------------------------------------------------- #!/usr/bin/perl -w use strict; die "Usage: rtdl [logfile]\n" unless @ARGV == 1; die "Unable to open pipe from tai64nlocal ){ my $ip = substr $_, 30, 8; my $x = substr $_, 39, 4; my $y = substr $_, 44, 4; if( $ip =~ /^([0-9a-f]{2})([0-9a-f]{2})([0-9a-f]{2})([0-9a-f]{2})$/ ){ $ip = sprintf( "%15s", (join '.', hex($1),hex($2),hex($3),hex($4))); } if( $x =~ /^([0-9a-f]{4})$/ ){ $x = sprintf "%05d", hex($1); } if( $y =~ /^([0-9a-f]{4})$/ ){ $y = sprintf "%05d", hex($1); } substr( $_, 30, 18 ) = join ':', $ip, $x, $y; print; } ------------------------------------------------------------------------ -- Aaron From nobody Fri Apr 4 12:18:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: The bane of my FreeBSD existance References: Organization: ESC Date: 04 Apr 2003 12:18:03 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 frank@SPAMLESSesperance-linux.co.uk (Frank Shute) writes: > Or if you're on unix when you edit the file, you should be able to > just open the file in vim, then do: > :w > and it should save it in unix format. It better not! I don't expect my editors to be altering files without my say-so. -- Aaron From nobody Sat Apr 5 14:35:06 2003 Newsgroups: comp.unix.bsd.freebsd.misc,comp.os.linux.misc Subject: Re: Is there a command for Unix which can detach process from console and wait until its finish? References: <169a7283.0304050443.30926141@posting.google.com> <169a7283.0304051031.14611908@posting.google.com> Organization: ESC Date: 05 Apr 2003 14:35:03 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 37 sdieselil@yahoo.com (sdieselil) writes: > YES. X thinks that _detach_ has exited and exits itself. So what I > need is that 'detach' will wait until icewm exits. But NOT waiting is kinda the definition of detach. If for some reason you want to detach icewm and run it in the background, make one of your xterms or something else run in the foreground. Basically, something that your .xinitrc starts needs to keep running in the foreground, and once that process exits, X will exit. Usually people choose to make their window manager that process, but it doesn't have to be. You could do: xterm & detach icewm exec xclock Then your X session would close whenever you close (or kill) xclock. > IceWM has a problem with my locale so it prints a warning and then > wait for TTY input (I checked it using running icewm from xterm). Sounds like a problem to fix. You can't really count on a window manager to run the same way from within an xterm that it does when started from .xinitrc. If you want to find out what problem it's having, put this in your .xinitrc: xterm & exec icewm >icewm.log 2>&1 That'll save whatever icewm is outputting to the file ~/icewm.log, so if it's waiting on something, we can figure out what it is. -- Aaron From nobody Sat Apr 5 17:44:16 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Lost login?!?!?! References: <3e8f4571_5@corp.newsgroups.com> Organization: ESC Date: 05 Apr 2003 17:44:16 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "Maje" writes: > Last night I compliled a kernel in order to add IPDIVERT and when I > was done...I was not able to log in using my username...I became > very concerned when I also could not log in using root. Then I > discovered that if I used root as a login and left the password > blank...I got right in!!! My first guess would be that when you ran mergemaster, you let it overwrite your /etc/master.passwd with the new one from the build. -- Aaron From nobody Sun Apr 6 11:09:49 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Lost login?!?!?! Organization: ESC Date: 06 Apr 2003 11:09:49 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 bv@wjv.comREMOVE (Bill Vermillion) writes: > Hm. I just ran mergemaster on a 4.8 upgrade, and your message > made me realize that I wasn't asked about the password files. > In the past there was always the new distribution file and the old > file, but this time the questions never came up. I haven't done any 4.8 upgrades yet, but I'll have to watch for that. > So - has this been changed in the 4.8? If so, that is a good thing > - as it has caused grief for more than 1 peron. I say that from > the above and one of my experiences in the past so that makes 2! But you always back up your /etc first, right? :-) -- Aaron From nobody Sun Apr 6 11:16:36 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Why generic kernel? References: <3E8ECE43.84BFFAFF@spam.de> <3E90371F.F8270AF3@spam.de> Organization: ESC Date: 06 Apr 2003 11:16:36 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 "R. Readi" writes: > Perhaps the installation programm could offer a minimal and a > generic kernel. It couldn't offer *a* minimal kernel; it would have to offer a variety of them. Once you start stripping the kernel down to get it to fit into smaller memory or disk, you're going to have to offer more than one, so people with different hardware can find one to use. That's how Slackware Linux has always done it. Pick an install floppy from a list, based on whether you're installing over a network, whether you have SCSI and what kind, and so on. There's nothing wrong with that approach, but it's extra trouble for the distributors. Are there enough people out there these days with <12MB RAM to make it worth the trouble? I don't know. I suspect that most of the people who are still using hardware that out of date (like me) are also the type that are savvy enough to swap drives or move some RAM around to get the job done with the standard install. -- Aaron From nobody Sun Apr 6 14:06:16 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: The bane of my FreeBSD existance References: <47ak6b.pb.ln@bigun.home> Organization: ESC Date: 06 Apr 2003 14:06:16 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 frank@SPAMLESSesperance-linux.co.uk (Frank Shute) writes: > I suggested using vim to the original poster as it's nice to be able > to use the same editor under unix & windows. I'm surprised the emacs > users haven't chipped in with how to change CRLF in emacs....bound > to involve some hideous meta/ctl key bashing...;) Just a little. ESC-< (goes to top of file) ESC-% C-q C-m ! (does a search and replace) I use both vi and emacs a lot, and I find myself reaching for the ESC key much more often with vi than with emacs, especially when jumping around a file editing small chunks of stuff. Maybe I could remap the 'leave editing mode' key in vi to something closer on the keyboard? -- Aaron From nobody Sun Apr 6 16:46:12 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: The bane of my FreeBSD existance Organization: ESC Date: 06 Apr 2003 16:46:12 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 per@hedeland.org (Per Hedeland) writes: > Well, it used to be that "simple", but nowadays (since around emacs > v20 I think) I find that loading a file with CRLF line termination > makes emacs go into "DOS mode" (indicated on the mode line), not > just hiding the CRs in the display but also preventing their > deletion with the above. Strange; I've never had that happen. I'm using XEmacs 21.1.14 right now, but I used a 20.x version of Emacs a while back and didn't notice it then. I don't edit dos files very much, though. It looks like there's such a thing as dos-mode, but I don't have it on my system. You might check your .emacs file for any reference to it, and comment out that line if you find it. If you don't find it there, you could search your system for dos-mode.el and dos-mode.elc and rename them, and see if that stops it. -- Aaron From nobody Sun Apr 6 16:54:22 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: building port graphics/povray References: Organization: ESC Date: 06 Apr 2003 16:54:22 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 25 Peter Wichert writes: > I'm trying to build the povray-port, in an up-to-date /usr/ports > with FBSD 5.0. > But all I get is a checksum-error: > I tried some more downloads, from other sources, too. But always the > same error. So I don't believe that it is a real checksum > problem. Does anybody have an idea? If you look in the Makefile, there's a MAINTAINER line that tells you who's in charge of the port. You could write to that person and ask. Since the povray source files don't have versioning in the filenames, most likely the current version has been updated, and the port's checksum hasn't been updated to match it yet. I actually had the same problem the other day, but I think in my case, it was because the download failed to complete (slow connection here). I'm trying it again now. -- Aaron From nobody Mon Apr 7 06:42:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: The bane of my FreeBSD existance Organization: ESC Date: 07 Apr 2003 06:42:39 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 23 per@hedeland.org (Per Hedeland) writes: > No, that's old stuff - I used the term "DOS mode" loosely, it's not > really a mode in the traditional sense but apparently part of the > MULE stuff in recent emacsen - a property of a "coding system". That could explain it; I'm pretty sure I'm using the non-MULE version. > I don't have a dos-mode.el* anywhere, and even without any .emacs at > all, I get this with emacs from the "standard" FreeBSD pkgs, > e.g. emacs-20.7_1 on FBSD 4.5 and emacs-21.2_2 on 5.0. XEmacs may be > different. > It's easy enough to check for it, just "manually" create a "DOS > file" by typing C-q C-m at the end of each line, save it, kill the > buffer, and load the created file - tada, invisible CRs!:-) Yep, I still get the normal ^M's with non-MULE XEmacs 21.1.14. -- Aaron From nobody Mon Apr 7 11:00:59 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: The bane of my FreeBSD existance Organization: ESC Date: 07 Apr 2003 11:00:59 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 37 frank@SPAMLESSesperance-linux.co.uk (Frank Shute) writes: > I don't mind the ESC key as it's about the only key I could find > blindfold! True. But it requires moving all my left-hand fingers from their keys and bringing them back. That's a lot of time-consuming movement for such a common keystroke. > With vim it's fairly easy to maybe map a function key to ESC. eg. Put: > map! #5 > > in ~/.vimrc to map F5 That's cool, but the F-keys aren't much closer than ESC, and closer together, making it easier to hit the wrong one. > :help map > & > :help keycodes > might be useful to you. Thanks, I'll check that out. Maybe I can map the tilde key to it, or even that dumb Windows key. > With vanilla vi or ex you can't map the function keys AFAIK so you > might be best off rebinding the key in X and/or the console or even > better just using vim. Guess I've got a port to install and try out. -- Aaron From nobody Mon Apr 7 16:39:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: building port graphics/povray Organization: ESC Date: 07 Apr 2003 16:39:41 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Peter Wichert writes: > Yes, it looks like a mismatch between the distfile and the port > directory, because ignoring the checksum does not help. The building > then breaks at the next step. Same here; it fails on the patches. If the port maintainer doesn't respond, I suppose the next step is to inspect the patches and try to install them and build it manually. -- Aaron From nobody Wed Apr 9 06:01:52 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: 4.7r -> 4.8r References: <2qm69vsh4q48ns6861mj0o45kulm36e7ik@4ax.com> Organization: ESC Date: 09 Apr 2003 06:01:52 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 30 bv@wjv.comREMOVE (Bill Vermillion) writes: > I've not time it in a long time. But set it up to run in cron and > you can sleep while it's being done. Cron makes it doable even on > 33K dialups. Yep. I have a nightly cron job that pulls src, doc, and ports. I'd say it generally takes about 15 minutes or so, but I'm sleeping, so it doesn't matter. > >If you have a lot of packages installed, and many of them have > >been updated since 4.7-R, it would be a lot quicker to install > >new packages from CD than by downloading source and building > >from ports. > I never touch the ones I have installed. I run portsupgrade now > and then to see what needs to be upgraded, but most I just leave > alone. Last time I had to build or install all new ones it took > forever [well a long time then]. a pkg_info | wc shows I > have 339 installed. A couple of them took a long time to compile > - the LaTeK for example. It's the old "If ain't broke don't fix it" > approach. Same here. I upgrade a port if there's a security issue, or if a newer version has a feature I need. Otherwise I leave them alone. -- Aaron From nobody Thu Apr 17 07:38:58 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Bootmanager References: Organization: ESC Date: 17 Apr 2003 07:38:58 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 Bernt Nilsson writes: > When I boot my machine with an oldish version of Windows and FreeBSD > 5.0 on separate partitions, I am presented with the choice between > "DOS" and "Disk 1". I guess this is the FreeBSD Bootmanager. > Is it possible to change the strings "DOS" and "Disk 1" to something > more correct? If so, how? Yes, do a google search in this group for "edit boot prompt" for directions on how to do it. The main thing is you can't change the length of the strings. -- Aaron From nobody Mon Apr 21 18:48:53 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: fbsd5 on Asus a7a266 motherboard References: Organization: ESC Date: 21 Apr 2003 18:48:52 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "Dave" writes: > Has anyone got Freebsd 5.0 release going on an Asus a7a266 > motherboard? If so, what hurdles did you have to go through if any > to make it work. I'm using it on an A7N266-VM, if that's similar. Everything works except the integrated MCP network interface, so I just use a PCI one. I posted here the problems I had with the integrated NVidia stuff and the solutions I found, so you can google for them. -- Aaron From nobody Mon Apr 28 15:53:15 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: grepping through log files for a value References: <3ead5875$0$230@hades.is.co.za> Organization: ESC Date: 28 Apr 2003 15:53:14 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 44 "Rudi Ahlers" writes: > I would like to see from /var/log/ppp, when the last time was that > PPP connected, and when it disconnected after that. Basically, I > would like to see, at any given time, how long I've been connected > upto that time. I don't know PHP, but you could do something like this in a perl CGI, or convert this to PHP: ------------------------------------------------------------------- #!/usr/bin/perl -w use strict; use Date::Calc qw(:all); print "Content-type: text/html\n\n"; open IN, "/var/log/ppp.log" or die $!; my @l = grep /IPCP: myaddr/, ; # Get lines denoting PPP connects my @date = split /[ :]/, pop @l; # split last one close IN or die $!; my $sec = time - Date_to_Time( 2003, Decode_Month($date[0]), @date[1..4] ); my $hours = int( $sec/3600 ); # subtracted time from last connect $sec %= 3600; # from current time, splitting my $mins = int( $sec/60 ); # into readable HH/MM/SS $sec %= 60; printf "PPP has been connected for %0d hours %0d minutes and %0d seconds\n", $hours, $mins, $sec; ------------------------------------------------------------------- I arbitrarily picked a unique string ("IPCP: myaddr") that I thought would only be printed on a successful connect, and then pulled the date info from it, converted in to seconds with the Date::Calc module from CPAN, subtracted that from current time to get elapsed seconds, and then print that in a decent format. You'll have to make sure /var/log/ppp.log is readable by the web server. Since the ppp.log doesn't include the year, I hardcoded it to keep this short, so you'll want to fix that if you're actually using this. -- Aaron From nobody Fri May 2 03:56:50 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: porting little c progs References: Organization: ESC Date: 02 May 2003 03:56:50 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 Keve Nagy writes: > What I have difficulties to find the corresponding functions for is > delay() and kbhit(). In Borland's conio.h the function kbhit() > should return 0 if no keystroke is in the keyboard-buffer, and a > non-zero value if there are any keystrokes. For example by a short > line of: Sounds like getchar(). `man 3 getchar` > The other instruction I miss is delay, where for example delay(3000) > was waiting for 3000 milliseconds (AKA 3 seconds) and then the > program execution continued. sleep(int) `man 3 sleep` -- Aaron From nobody Fri May 2 14:35:34 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Need help with Installation References: <3eb1d01d$1@mtunews.mtu.edu> Organization: ESC Date: 02 May 2003 14:35:32 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 "alexwa" writes: > Ok, I have startx configured and kde opens with no problem. > However, my XF86config file is messed up as my screen resolution is > wrong, it should be 1024 x 768, not 800 x600. Also the screen is > off center to the right. Not sure what I did wrong. How do I fix > this? Edit /etc/X11/XF86Config, and look for the "Screen" Section. First, see what DefaultDepth is set to. Then look for the Subsection "Display" that has that Depth. This Subjection will have a Modes line that looks something like this: Modes "640x480" "800x600" "1024x768" "1280x1024" Those are the possible resolutions that you'll get, in that order. You can cycle through them one by one with Ctrl-Alt-Plus and Ctrl-Alt-Minus. To change which one you get by default, change the order so the one you want is first. If you want to change the color depth, change the DefaultDepth setting, and then make sure the Modes line in the corresponding Subsection is what you want. > Also I get a message saying my sound input is not configured. Any > ideas? You have to recompile your kernel to get sound support. See the handbook for instructions. If you've already done that, you might have kde issues, which I can't help with. -- Aaron From nobody Mon May 5 19:09:44 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: May 5 20:47:23 Valiant named[261]: denied update from [192.168.0.20].3812 for "Camelot.Castle" IN References: <3eb6d51c$0$236@hades.is.co.za> Organization: ESC Date: 05 May 2003 19:09:44 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "Rudi Ahlers" writes: > What does this mean? It means you've got a system on your local network that's trying to tell the DNS server what its name and address are, so the DNS can update its information. Probably a Windows system. You can just ignore it, and I think the named man page should tell you how to get it to stop logging it. -- Aaron From nobody Mon May 5 20:48:42 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Alternative to atoll() Organization: ESC Date: 05 May 2003 20:48:42 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 11 I'm trying to compile the latest CVS version of the game Crossfire on FreeBSD 4.7. It uses the atoll() function in a couple places. Apparently this is similar to atoi() and atol() except that it returns a 64-bit value. I don't appear to have this function on my system. Anyone know of an alternative? I'm not a great C programmer, obviously. -- Aaron From nobody Tue May 6 06:09:02 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Alternative to atoll() Organization: ESC Date: 06 May 2003 06:09:02 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 charon writes: > from man atol > atol() is equivalent to strtol(nptr, (char **)NULL, 10); > > so I guess atoll would be equivalent to some permutation of > strtoll (see man stroll) Thanks, I think that'll do it! -- Aaron From nobody Tue May 6 06:25:09 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: network dies without a reason References: <20030505221057.55b1ca31.tsprad+usenet@spradley.org.invalid> Organization: ESC Date: 06 May 2003 06:25:09 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Ted Spradley writes: > I used to have a PNIC ethernet board that would sometimes > spontaneously get wedged. ifconfig down and back up would fix it. > Sometimes it would be OK for two or three months, sometimes it would > do it twice in 24 hours. I've never seen it happen on any other of > the cheap junk ethernet boards I've used. I've had that happen a couple times. In one case it was a bad network card; in the other case new a new cable fixed it. -- Aaron From nobody Tue May 6 07:56:00 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: partioning References: Organization: ESC Date: 06 May 2003 07:56:00 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 Jeff Davis writes: > If I'm going to run a mail server, would I want to allocate a lot of > space to /var? I would imagine a few hundred pop3 accounts eat > through 2GB fast. Depends on your mail server. I use qmail, which stores mail in the user's home directory. I prefer that because it keeps all their files in one place, which makes backups and controlling their usage easier. > It would also seem to me to give more space to /var because of log > files. I'm new to FreeBSD so I'm not really sure if logrotate or > something moves them somewhere else. FreeBSD has newsyslog, which has a simpler setup than logrotate. I have servers with /var as small as 20MB, and newsyslog keeps them cleaned up nicely. The main other thing that might take up space on /var is a database. I prefer to put this on my big partition (/usr or /home) and link to it. MySQL, for example, wants to find the database files in /var/db/mysql if you install from the port, so I link that to /home/mysql or somewhere. This way, my /var partition is strictly dynamic stuff like log files that I can live without in case of data loss, so I don't have to backup that partition as often if I don't want to. It also means that if a log file grows out of control and fills the partition, important stuff like mail and database writes can still be done on the other partitions. -- Aaron From nobody Thu May 8 06:46:22 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: partioning Organization: ESC Date: 08 May 2003 06:46:22 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 31 charon writes: > Aaron Baugher wrote: > > The main other thing that might take up space on /var is a > > database. I prefer to put this on my big partition (/usr or > > /home) and link to it. MySQL, for example, wants to find the > > database files in /var/db/mysql if you install from the port, so I > > link that to /home/mysql or somewhere. > no need to link anymore - theres > DB_DIR Set alternate directory for database files in the > Makefile. 'e.g. make DB_DIR=/home/mysql/data' True, but then you have to remember to do that every time you update the port. (Does portupgrade keep track of such options and apply them on future upgrades?) By using the link, I don't have to worry about forgetting to add that option, especially on systems where I'm not the only person with permission to upgrade ports. Also, I've put the real directory in different places on different clients' systems thanks to different partitioning schemes; but I know I can always cd to /var/db/mysql to find it. I suppose I could set DB_DIR in make.conf, but that assumes that no other port uses the same variable. Too bad it's not something more unique and memorable, like MYSQL_DB_DIR. -- Aaron From nobody Thu May 8 07:00:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Router configuration References: <20030508024425.00001356.nongamorc@hotmail.com> <20030508085431.29398c53.einherjer01@hotmail-spammershaven.com> Organization: ESC Date: 08 May 2003 07:00:40 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 Robert writes: > If you use sendmail_enable="NONE" remember that the cron jobs > produce every night a email. Which is now only written in the > /var/spool/mqueue and never send out. So your var disk could run out > of disc space. I don't believe so. The sendmail program (not the daemon) will still put local mail in /var/mail and send remote mail off the system. Outgoing mail that doesn't make it on the first try will be queued, but /etc/periodic/daily/500.queuerun will retry any queued messages by running 'sendmail -q' once daily. That'll keep the queue tidy, saving only the messages that are younger than whatever you have the queue timeout set to. -- Aaron From nobody Sat May 10 19:10:09 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Router configuration Organization: ESC Date: 10 May 2003 19:10:08 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 per@hedeland.org (Per Hedeland) writes: > Not so - as of sendmail 8.12 (FreeBSD 4.6 I believe) you need the > daemon running and listening at least to the loopback interface to > get any mail delivery (local or outgoing) to happen. The mail > submitted by direct invocation of sendmail will actually get stuck > in /var/spool/clientmqueue rather than /var/spool/mqueue, though. All I know is it's working fine for me on a 4.7-RELEASE system, with absolutely no sendmail daemon running and nothing listening on port 25. Local mail is being delivered to /var/mail and outgoing mail sent with /usr/bin/mail reaches its destination. -- Aaron From nobody Sun May 11 07:41:23 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD is not the answer References: <20030510220513.22341eb6.tsprad+usenet@spradley.org.invalid> Organization: ESC Date: 11 May 2003 07:41:21 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 85 Ted Spradley writes: > I see this sort of thing often, and I'm always amazed. It's just so > completely opposite my own experience. Unix came to me so easily > and naturally, right from the beginning. Unix (and especially BSD > Unix) has always seemed easier to learn and understand than any > other system I've ever tried to use, before or since. Windows, on > the other hand, continues, even after all these years, to be > obscure, clumsy to use, incomprehensible to understand, and > incapable of being coerced into doing any sort of useful work > without significant expenditures of not just effort but especially > of money. > Is my experience really so unique? [Didn't intend for this to turn into a long rant; it just got away from me.] No, that was my experience too. I'd guess it depends greatly on which system you got comfortable with first. I spent a year using CP/M and programming in assembly, then got a Commodore 128 which had BASIC and machine code for "languages," and then got an Internet account with an ISP that offered shell accounts on a Unix system (Sun, I think). My only experience with MS-DOS was playing a few games on a friend's PC. Sure, the graphics were nice, but the system didn't seem useful for much else. Using that Sun, the powerfulness of the Unix way of doing things was obvious. Before long I was writing shell scripts, building long pipelines of various commands like grep and sort, and using telnet, ftp, gopher, elm, and tin to access the Internet. I even did some C compiling, to compile a program used to transfer files between that system and my C128. When I later started working for an ISP, I had to learn enough about Windows to help people get their modems working and their Internet software installed. (Trumpet Winsock on Win3.1, at the time.) The primitiveness of Windows and DOS compared to what I'd gotten used to with Unix was astounding. BAT files? I had a handful of shell languages that were a hundred times more useful. Lettered drives? Please. Useful tools? I had better tools like grep on my 2Mhz C128 running Craig Bruce's ACE shell. I immediately realized Windows was a toy system, perhaps useful for playing games or doing some very simplistic home computing, but certainly not for anyone who liked computers or wanted to get the most out of them. While Windows has improved, so has Unix, so the comparison still applies. If most kids' first computer usage over the last couple decades had been with Unix systems, MS-DOS and Windows would be about as common now as CP/M and OS/2. I can understand where people are coming from with the "it's too hard to learn" complaint, though. With Unix, there's so much power under the hood that it encourages -- or even demands -- that you take advantage of it. It's like owning a muscle car with 500 horsepower and a top speed of 150 mph. You're not going to be satisfied tooling it around at 55; sooner or later you're going to have to see what it can do. Then you're going to have to pop the hood and see what's under there giving you all that power, and you'll probably want to tinker with it to make it even better. Windows users, for the most part, don't tinker. (When their kids tinker, they end up calling tech support to get help fixing it.) The OS doesn't give them any tools to do any useful tinkering anyway. As long as their computer does more or less what it did when they bought it, like an appliance, they're happy with it. When those folks are drawn to Unix, they may want to tinker a little -- new tires, maybe change their own oil -- but they don't want to rebuild the engine to squeeze out more horses. No Unix system draws a sharp line between 'using' and 'tinkering'; once you login to a shell, you've got all the tools right there at your fingertips. Personally, I don't have a problem with that. I've never understood the goal of making computers so easy to use that anyone can use them without any instruction at all. We don't take that attitude about other tools like guns, cars, or washing machines, even though those are all much simpler than any computer operating system. Would it really be so terrible if people had to learn a few things to use a computer, and then a few more things to get online? -- Aaron From nobody Mon May 12 06:06:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD is not the answer Organization: ESC Date: 12 May 2003 06:06:13 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 29 talon@lpthe.jussieu.fr writes: > In fact Windows is becoming tremendously better than it was in the > Win98 days. I have installed WinXP Pro, and to my great surprise, > things are not terribly different as on a Unix machine. To have the > pleasure of writing shell scripts at the console, i have installed > cygwin, which works *very* well. So if I install a third-party program that makes Windows act like Unix, that makes Windows a better OS. Makes sense to me. > In order to be able to browse web and use mail and news exactly as > on a Unix machine i have installed Mozilla, and to read ugly .doc > documents i have installed OpenOffice. With such setup, things are > so similar under Windows as, say , FreeBSD that it is hard to see > great difference. Multitasking works correctly, I agree that XP is a big improvement over previous Windows incarnations, but I've still had program lockups that hosed the entire machine and required a reboot. There's still too much "cooperation" between the programs and the OS for me to compare it favorably to Unix. On my FreeBSD machine, I never worry that opening a twentieth application is going to hose any of the other nineteen that are already running. -- Aaron From nobody Mon May 12 06:18:57 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD misses the boat (well, sorta) References: Organization: ESC Date: 12 May 2003 06:18:57 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 Joe Potter writes: > I notice that FreeBSD is most like Linux Slackware in lack of GUI > help thingys that RedHat has. Yes, that's one of my favorite things about it. My one Linux machine is Slackware. It was the first Linux distro I ever used, and I always come back to it after trying the flashier ones and being disappointed. > The real place where FreeBSD misses the boat is not letting folks > know how easy it is to use the ports. I agree. The first few times I installed FreeBSD I just naturally started installing everything else from source the way I did on Slackware -- download, untar, ./configure && make && make install. Somewhere along the line I discovered the ports tree, and realized how much unnecessary extra work I'd been doing. The ports tree definitely deserved more than an off-hand mention in the middle of sysinstall. Every newbie to FreeBSD should be guided to the ports as strongly as possible. -- Aaron From nobody Mon May 12 06:24:06 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Router configuration Organization: ESC Date: 12 May 2003 06:24:06 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 per@hedeland.org (Per Hedeland) writes: > Well, it's possible to run sendmail 8.12 the "oldfashioned" way, > without /etc/mail/submit.cf and with the binary setuid-root, but > that's not the default setup on 4.7-RELEASE as far as I remember > (and the files I saved before upgrading from it seem to agree). I'm running 4.7-RELEASE-p2, and that's still my setup. I've never done any sendmail configuration on this machine other than SENDMAIL=NONE, and I let mergemaster replace everything in /etc/mail that it wants to. Thanks for the heads-up; I'll remember to watch for this next time I upgrade that machine. Normally one of the first things I do when I setup a server is to kill everything sendmail and install qmail, but this machine didn't need to accept incoming mail, so I didn't bother. -- Aaron From nobody Mon May 12 23:00:18 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD is not the answer Organization: ESC Date: 12 May 2003 23:00:18 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 William Clardy writes: > Donn Miller wrote: > >Ah, but what can Windows XP do that FreeBSD or Linux cannot? > Run Java 1.4.1 hassle-free. Hmm. When I installed XP Pro, I didn't have a Java VM. I went hunting at Microsoft's site, and it appeared that they weren't offering it anymore. I installed Sun's, and it sort of worked, but applets ran about 1/20 the speed that they had under Win98. So I went hunting on newsgroups, and eventually found detailed instructions for manually installing a certain XP Update that included Microsoft's Java VM. Problem fixed, but definitely not something I'd expect an average user to be able to figure out. -- Aaron From nobody Tue May 13 06:35:49 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newsgroup hosting References: Organization: ESC Date: 13 May 2003 06:35:47 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 40 flippytheclown writes: > Currently the company that I work owns an ISP with <2,000 dialup > customers as well as a dozen fractional T1 commercial accounts. We > currently pay Supernews.com about $200/month to host up to 10 > concurrent sessions to browse newsgroups. I think this might be a > good opportunity to have a FreeBSD server do this instead. > I am not sure where to start. Here is a short list of > questions/comments that I have. > 1) What software is out there and what is the prevailing/best > choice. I found INN in ports collection is this a good place to > start. I use leafnode, which works great for my one-person news server. It's a dynamic server, meaning that it just gets the groups that users' newsreaders have requested. Leafnode is recommended for light duty; there are other heavier-duty servers that act the same way, like dnews. I haven't used INN in several years, so I don't know if it'll do this now or not; when I used it it would only accept full feeds. > 4) Traffic on our hosted connections are relatively light seldom > more than a few concurrent connections. Idea on how big the server > would have to be or what factors I should consider in sizing one. It really depends on what newsgroups your users hit. Do you plan to offer the binaries groups? Some of the binaries groups where people post music and movie files are *huge*. If it's possible to get a report from Supernews that shows what newsgroups your users are reading, that would help you plan for the bandwidth and space necessary. If you want to block binaries groups, I think almost any server package will do that. -- Aaron From nobody Tue May 13 19:14:51 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newsgroup hosting Organization: ESC Date: 13 May 2003 19:14:50 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 25 flippytheclown writes: > Thanks for your help. I checked out their website > http://www.leafnode.org looks like it's worth a looksie. I notice > in the ports collection there are 2 ports leafnode and leafnode+. Do > you know what the difference is? As I understand it, when leafnode reached version 1.4, the original authors stopped development on it. Someone wanted to add more features that hadn't been implemented yet, like the ability to suck news from multiple servers, so he started a new branch from his own patches and called it leafnode+. Development also started again on the original leafnode, which is now at version 1.9.40. Functionally, I think they're very similar. Both will get news from multiple servers, and the main programs work similarly. Leafnode has the ability to filter newsgroups based on regular expression tests on the headers; I don't know if leafnode+ offers that. I've always used and been happy with Leafnode, the original, so I can't really say how leafnode+ compares. -- Aaron From nobody Wed May 14 07:08:04 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: hd duplication with dd References: Organization: ESC Date: 14 May 2003 07:08:03 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 44 Keve Nagy writes: > I use dd to make exact copies of the ad0s2 and ad0 (ad0s3 and ad1 is > never mounted) > My only problem is that a weekly dd if=/dev/ad0 of=/dev/ad1 bs=512 > takes more than 17 hours. In a situation like that, I use rsync. It only copies the files that have changed since the last backup, so it should be much faster than copying the entire drive. The only problem with rsync is that it builds a list of files before doing the copy, so you can run into RAM problems if that list gets extremely long. If that's a problem, I'll use a script like the one below to split the job up into sections and skip the partitions that I don't bother with, like /proc and /tmp. It also mounts and unmounts the destination drive, so no-one can mess with it the rest of the time. ---------------------------------------------------------------- #!/bin/sh mount $DEST_DRIVE /old if [ -f /old/.unmounted ] ; then echo Backup disk not mounted; quitting. else cd / for i in bin boot command etc home lib opt package root \ sbin service usr var ; do echo Backing up $i ..... nice rsync -v -r -l -p -o -g -t --delete $i /old done umount /old fi ---------------------------------------------------------------- The file .unmounted exists in the otherwise empty /old directory, but not on the mounted drive's /. That way I can test for its existence; if it exists after the mount, the mount failed. -- Aaron From nobody Fri May 16 06:41:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Favorite X setup? References: Organization: ESC Date: 16 May 2003 06:41:11 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 Gary Schenk writes: > Soon I'm going to set up my final (for awhile) install of > FreeBSD. I've been using KDE up until now, but I'm considering > running windowmaker, or icewm, with the idea of getting my machine > as lean as possible. It's a P3 600MHz, 128 MB RAM, 20 GB HD. > Does anyone have a favorite combination of X software? The things I run all the time are windowmaker, lots of basic xterms, galeon, and XEmacs. Those aren't all the lightest-weight options, but I ran them all on a 200Mhz Pentium without too much slowdown. > What to use for burning CDs? My CD is not SCSI. Use burncd. After messing around with xcdroast on Linux (needing SCSI emulation), I was amazed that FreeBSD had something so simple and easy to use -- and right in the base system, no less. > What would you consider a good replacement for kmail, knode, kscd, > and konqueror? Well, you don't have to replace any of them just to change window managers. You can keep using all of those with any other window manager. I prefer galeon to konqueror, but it is more demanding on the system. Every web browser I've ever used crashes once in a while, so galeon's ability to restart with all the same settings and destinations intact is crucial. A really lightweight browser option would be dillo, but it's missing some basic features I can't do without. I use XEmacs for most file editing and programming and gnus within XEmacs for mail and news, so I get a lot of productivity out of the large 30-40MB footprint it takes. I don't know what knode and kscd are. -- Aaron From nobody Fri May 16 06:47:49 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: maildrop port + mysql References: <65of98cu.fsf@6e65747363617065.6e6574> Organization: ESC Date: 16 May 2003 06:47:49 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 Matt M <6e777a6d617474@6e65747363617065.6e6574> writes: > I know. I had mysql-client installed, which is why I didn't > understand what was going on. I am still fairly new to FreeBSd and > the ports tree. I also don't fully understand the default > environment that `make' uses and where it is set. It turnes out that > libmysqlclient.* are located in /usr/local/lib/. For some reason > that is not part of the LIB path for make. I ended up creating > symlinks to them into /usr/lib/ and everything ran smooth from > there. When you install the mysql-client port, it creates the file /usr/local/etc/rc.d/mysql-client.sh (or similar). This runs the command: /sbin/ldconfig -m /usr/local/lib/mysql Which merges that directory into the list of shared libraries to be used. That command should also be run when the port finishes installing, but you might want to run it yourself to see if that helps. -- Aaron From nobody Fri May 16 19:28:28 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Cannot access network device rl0 References: Organization: ESC Date: 16 May 2003 19:28:28 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 "Mike Hunt" writes: > I'm a FreeBSD newbie who's been using Linux for awhile. I am using > FreeBSD-4.8 STABLE. Unfortuanely, I cannot get the network device > rl0 to work. It is not listed in ifconfig -a and during boot I get > the following messages: Realtek NICs are supported in the generic kernel, so if yours doesn't work by default, it's probably not going to. The consensus seems to be that they're very low-quality cards anyway. If yours is integrated into the motherboard, which seems to be the case with many Realteks that people have trouble with, I'd suggest turning it off in the BIOS and PCI network card. -- Aaron From nobody Fri May 23 06:58:50 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: broken ports, or my own mistake? References: Organization: ESC Date: 23 May 2003 06:58:48 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 52 Ben Crowell writes: > Here's a synopsis of each problem. After that, I've posted the > actual error messages. These problems all occurred after doing a > fresh, complete cvsup this morning. > 1. building net/librsync > This failed because it needed /usr/local/bin/sed_inplace, which is > in the textproc/sed_inplace port. Does this indicate an error in the > librsync port, which should automatically have known how to satisfy > this dependency? Probably, but it could be that sed_inplace exists in the base system in FreeBSD 5.x, so the port maintainer didn't need a dependency. At least it makes it obvious what you need to do -- just install the sed_inplace port and then come back to this one. > 2. building devel/popt > This failed with the message "Patch patch-aa failed to apply cleanly." This could mean that you happened to cvsup the port right as the maintainer was updating it, and got a distfile that didn't match the patches. I just installed it, and it went fine. The first thing I'd try would be to do a new cvsup and try building it again. > 3. building bison > This failed as it was attempting to make the texinfo > documentation. The binary package seemed to be OK. That rings a bell, like I've had that problem before, but I don't remember the details. If you post the last several lines of output, we might be able to see what's wrong. Or you could edit Makefile, and try to get it to skip the texinfo part. > makeinfo --no-split -I . `test -f 'bison.texinfo' || echo > './'`bison.texinfo -o bison.info > bison.texinfo:37: Unknown command `copying'. > bison.texinfo:58: Unmatched `@end'. > bison.texinfo:93: Unknown command `insertcopying'. You didn't mention what version of FreeBSD you're using, but makeinfo is part of the base distribution. My 4.7 and 4.8 systems have version 4.2 of makeinfo. I'd say either your makeinfo is old or that file (bison.texinfo) has errors. -- Aaron From nobody Fri May 23 07:11:25 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Port Blocking by ISP References: Organization: ESC Date: 23 May 2003 07:11:24 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 "Max Snell" writes: > I have cable modem service by RCN in New York City. After some > tweaking I finally have my FreeBSD-4.6 server running as a gateway > for my LAN and I can share the connection with other computers. > In testing the capabilities of my new setup I tried sending some > mail from the FreeBSD server. Of course it did not work which means > the ISP is probably blocking port 25. Not too likely. Outgoing mail from your system will connect from some port above 1024 to port 25 on the destination server. It's possible that your ISP is blocking outgoing traffic to port 25, but that would be unusual. Normally they only block incoming traffic to prevent you from running servers. Try this: telnet mx1.mail.yahoo.com 25 You should get a message about SMTP service ready. If you don't, then either your ISP is blocking you or you're blocking yourself somehow, possibly with your firewall, if you have one. If you do get a ready message, then your problem is probably within your mail server configuration. If they are blocking outgoing->25, you should be able to tell your mail server to relay all outgoing mail to your ISPs mail server, and let theirs take it from there. I don't know how you do that with sendmail, but someone will surely chime in with the answer. -- Aaron From nobody Sun May 25 07:24:51 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Resizing partition References: <1hvkab.qjb.ln@xor> <7k8f8qx6.fsf@6e65747363617065.6e6574> Organization: ESC Date: 25 May 2003 07:24:51 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 21 Matt M <6e777a6d617474@6e65747363617065.6e6574> writes: > - use tar to copy all the files from /usr > > # cd / > # tar cfv /mnt/usr/usr.tar usr > # cd /mnt/usr > # tar xfv usr.tar In case you don't have room for a tar file, you can pipe tar to tar: ( cd / && tar -cf - usr ) | ( cd /mnt/usr && tar -xpvf - ) The 'cd &&' method makes sure that if you aren't able to cd to those directories, the tars never happen, so you won't accidentally untar somewhere you didn't mean to. -- Aaron From nobody Sun May 25 07:27:07 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Duke Nukem 3D Organization: ESC Date: 25 May 2003 07:27:06 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 12 I just discovered that the source code for Duke Nukem has been released, and there's a functioning Linux port out for it now. Is anyone working on a port to FreeBSD? It's probably beyond my abilities, but I'm grabbing the Linux code by CVS to take a look at it. If someone's already started this project, I'd like to know. -- Aaron From nobody Sun May 25 09:53:12 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Duke Nukem 3D Organization: ESC Date: 25 May 2003 09:53:12 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 34 > I just discovered that the source code for Duke Nukem has been > released, and there's a functioning Linux port out for it now. Is > anyone working on a port to FreeBSD? It's probably beyond my > abilities, but I'm grabbing the Linux code by CVS to take a look at > it. If someone's already started this project, I'd like to know. To followup on my own post: I CVS'd the source and built it with amazingly few changes. Just had to use gmake instead of make, and turn on USE_I386_ASM in the buildengine Makefile. There were compiler warnings, but it finished and the executable runs. Now for the bad news -- it's extremely slow, and lowering the detail in the game doesn't make any difference, so I don't think it's really a resource problem. The Linux port page doesn't say anything about it being unplayably slow, so I think this must be an issue with porting it to FreeBSD. I'll have to dig into it and see what I can find. Can anyone recommend any good resources for porting Linux apps to FreeBSD? Also, the sound has problems. Some sounds, like water dropping, are okay, while gunfire just turns into a long burst of static. Still, this is as close as I've gotten to having Duke3D running on my system (after much hopeless fiddling with dos emulators and the like), so I'm pretty excited. -- Aaron From nobody Sun May 25 16:34:27 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Duke Nukem 3D Organization: ESC Date: 25 May 2003 16:34:24 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 50 David Vidal Rodríguez writes: > Duke Nukem 3D on FreeBSD? Yes, it did compile! > However, I had to do some workarouds in the makefiles and #include > sections. For today's CVS-version, I have done a little patch, which > can be downloaded here: Great, that cleans up the compiler warnings. > In source/buildengine, you have to compile a.c manually and assemble > a_nasm.asm like this: > nasm -f elf -o a_nasm.o a_nasm.asm If I do that, I end up with an executable that won't run and dies with the error: Fatal signal: Bus Error (SDL Parachute Deployed) If I instead uncomment this line in source/buildengine/Makefile before making the buildengine: USE_ASM := -DUSE_I386_ASM I get an executable that runs and seems to work fine in every way except that it's extremely slow. (Like nearly a full second between shots with the pistol slow.) > The executable is built, however, I am having problems with the .CON > files. I don't seem to have any problem with those. I did fix my sound problems by editing DUKE3D.CFG like this: [Sound Setup] FXDevice = 6 MusicDevice = 6 FXVolume = 220 MusicVolume = 136 NumVoices = 8 NumChannels = 2 NumBits = 16 MixRate = 11025 Got that off the forum at 3dRealms. Sound is fine now. -- Aaron From nobody Wed May 28 06:46:16 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Profiling with gcc Organization: ESC Date: 28 May 2003 06:46:16 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 30 In my quest to port Duke Nukem 3D to FreeBSD, I'm trying to profile the program to see what's making it so low, as suggested on the 3drealms forum. I add -pg to CFLAGS and LDFLAGS, and when it gets to the linking stage, it dies like this: gcc -o build -g -pg -L/usr/local/lib -Wl,-rpath,/usr/local/lib \ -lSDL-1.1 -pthread build.o bstub.o engine.o cache1d.o \ sdl_driver.o unix_compat.o a.o pragmas.o /usr/local/lib/libaa.so.1: undefined reference to `longjmp' /usr/local/lib/libaa.so.1: undefined reference to `setjmp' /usr/X11R6/lib/libX11.so.6: undefined reference to `getgid' /usr/local/lib/libaa.so.1: undefined reference to `random' /usr/local/lib/libesd.so.2: undefined reference to `connect' [snip about 50 more similar lines] If I take the -pg out, it compiles fine again. After googling, I thought I might be missing the profiling libraries, so I set NOPROFILE to false in /etc/make.conf and did a full installworld and kernel (4.8-R). But I still get the same error above. I looked in /usr/lib, and I do have a bunch of lib*_p.a files, like libc_p.a. They're all dated from the day I originally installed this system, so it doesn't look like my installworld installed them, but they do exist. Anyone know what I'm missing here? Thanks, -- Aaron From nobody Wed May 28 06:58:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Nvidia binary driver issue References: Organization: ESC Date: 28 May 2003 06:58:11 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 27 Devon Ryan writes: > It actually happened to be some weirdness in my bios that I had to > correct (followed by an IRQ conflict that is now resolved). So now > the card works...but the driver is freezing the system (the vertical > lines problem) when I startx. This work fine if I use the nv > driver...but this is not optimal. Since this is a PCI card, what > should I set NvAgp to and does anybody know if loading the agp > kernel module is still required (I think I've tied every iteration > on loading/unloading agp with each NvAgp option). X freezing the > system is kind of annoying. Mine's integrated into the MB, but it sounds like the same problem. I don't load AGP; my /boot/loader.conf just has: nvidia_load="YES" Installing the driver from the port didn't work for me. I had to download the driver (NVIDIA_FreeBSD-1.0-3203.tar.gz at the time) and install it manually, following the instructions that came with it. I don't remember the details, but if you search this group for my address and NVIDIA, you should find my posts explaining how it went. -- Aaron From nobody Wed May 28 08:12:41 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Profiling with gcc Organization: ESC Date: 28 May 2003 08:12:39 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 Gunther Nikl writes: > > gcc -o build -g -pg -L/usr/local/lib -Wl,-rpath,/usr/local/lib \ > > Is -L<> -Wl,<> necessary? I don't know, but it works fine with them in there as long as I remove -pg. I'll try it without. .... Nope, no luck. Removing -W<> still gives the same errors. Removing -L<> makes it unable to find the SDL libs at all. > Did you add NOPROFILE=false? IMHO, NOPROFILE shouldn't be there at > all to get the profiled libraries. Yes, I did, because I wasn't sure if the default was true or false. I'm not finding anything under /usr/obj that looks like profiling libraries, though, except for lib/libc/profile.* and lib/libc_r/profil.*. Maybe my NOPROFILE setting confused it; I'll try rebuilding again with that removed. Thanks, -- Aaron From nobody Wed May 28 13:18:11 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Profiling with gcc Organization: ESC Date: 28 May 2003 13:18:11 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 40 > Yes, I did, because I wasn't sure if the default was true or false. > I'm not finding anything under /usr/obj that looks like profiling > libraries, though, except for lib/libc/profile.* and > lib/libc_r/profil.*. Maybe my NOPROFILE setting confused it; I'll try > rebuilding again with that removed. No such luck. I built and installed world without a NOPROFILE line in /etc/make.conf, confirmed that it copied a bunch of lib*_p.a files into /usr/lib, and rebooted. I still get the same 'undefined reference' errors when I give gcc the -pg option. I wrote a simple hello.c program and compiled it with -pg, ran it, and ran gprof on the resulting .gmon file. So profiling does work in general; it just won't with this command line: gcc -o build -g -pg -L/usr/local/lib -Wl,-rpath,/usr/local/lib \ -lSDL-1.1 -pthread build.o bstub.o engine.o cache1d.o \ sdl_driver.o unix_compat.o a.o pragmas.o /usr/local/lib/libaa.so.1: undefined reference to `longjmp' /usr/local/lib/libaa.so.1: undefined reference to `setjmp' /usr/X11R6/lib/libX11.so.6: undefined reference to `getgid' /usr/local/lib/libaa.so.1: undefined reference to `random' /usr/local/lib/libesd.so.2: undefined reference to `connect' [and so on] I also tried installing the ngpt port, to see if that implementation of pthread would help. No change. I think I've narrowed the problem down to a conflict between profiling and either SDL or pthread (which SDL requires). I've exhausted all the leads I could find through google searches. Right now I'm installing lang/gcc33, to give that a try. Any other ideas? -- Aaron From nobody Wed May 28 16:24:12 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Profiling with gcc Organization: ESC Date: 28 May 2003 16:24:12 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 120 kargl@troutmask.apl.washington.edu (Steven G. Kargl) writes: > > gcc -o build -g -pg -L/usr/local/lib -Wl,-rpath,/usr/local/lib \ > > -lSDL-1.1 -pthread build.o bstub.o engine.o cache1d.o \ > > sdl_driver.o unix_compat.o a.o pragmas.o > > Is something missing here? Nope, that's cut-and-pasted from the output; I just split the line. > Add -v to the command line and post the entire output. Here goes (splitting long lines): gcc -o build -v -g -pg -L/usr/local/lib -Wl,-rpath,/usr/local/lib -lSDL-1.1 -pthread build.o bstub.o engine.o cache1d.o sdl_driver.o unix_compat.o a.o pragmas.o Using builtin specs. gcc version 2.95.4 20020320 [FreeBSD] /usr/libexec/elf/ld -V -dynamic-linker /usr/libexec/ld-elf.so.1 -o build /usr/lib/gcrt1.o /usr/lib/crti.o /usr/lib/crtbegin.o -L/usr/local/lib -L/usr/lib -rpath /usr/local/lib -lSDL-1.1 build.o bstub.o engine.o cache1d.o sdl_driver.o unix_compat.o a.o pragmas.o -lgcc_p -lc_r_p -lgcc_p /usr/lib/crtend.o /usr/lib/crtn.o GNU ld version 2.12.1 [FreeBSD] 2002-07-20 Supported emulations: elf_i386 /usr/local/lib/libaa.so.1: undefined reference to `longjmp' /usr/local/lib/libaa.so.1: undefined reference to `setjmp' /usr/X11R6/lib/libX11.so.6: undefined reference to `getgid' /usr/local/lib/libaa.so.1: undefined reference to `random' /usr/local/lib/libesd.so.2: undefined reference to `connect' /usr/X11R6/lib/libX11.so.6: undefined reference to `xdrmem_create' /usr/local/lib/libaa.so.1: undefined reference to `atol' /usr/X11R6/lib/libX11.so.6: undefined reference to `getegid' /usr/X11R6/lib/libX11.so.6: undefined reference to `__inet_addr' /usr/local/lib/libvga.so.1: undefined reference to `usleep' /usr/lib/libncurses.so.5: undefined reference to `vsscanf' /usr/lib/libm.so.2: undefined reference to `__infinity' /usr/local/lib/libesd.so.2: undefined reference to `execl' /usr/local/lib/libvga.so.1: undefined reference to `puts' /usr/local/lib/libesd.so.2: undefined reference to `getuid' /usr/local/lib/libvga.so.1: undefined reference to `system' /usr/lib/libncurses.so.5: undefined reference to `remove' /usr/local/lib/libesd.so.2: undefined reference to `endpwent' /usr/X11R6/lib/libX11.so.6: undefined reference to `sleep' /usr/local/lib/libaudiofile.so.0: undefined reference to `ldexp' /usr/local/lib/libaudiofile.so.0: undefined reference to `vsnprintf' /usr/X11R6/lib/libX11.so.6: undefined reference to `authdes_create' /usr/X11R6/lib/libX11.so.6: undefined reference to `wctomb' /usr/local/lib/libesd.so.2: undefined reference to `socket' /usr/X11R6/lib/libX11.so.6: undefined reference to `mbstowcs' /usr/X11R6/lib/libX11.so.6: undefined reference to `strncasecmp' /usr/local/lib/libvga.so.1: undefined reference to `setpgid' /usr/local/lib/libesd.so.2: undefined reference to `pipe' /usr/lib/libncurses.so.5: undefined reference to `getpgrp' /usr/local/lib/libartsc.so.0: undefined reference to `strrchr' /usr/X11R6/lib/libX11.so.6: undefined reference to `readv' /usr/local/lib/libvga.so.1: undefined reference to `vprintf' /usr/local/lib/libesd.so.2: undefined reference to `__inet_aton' /usr/local/lib/libesd.so.2: undefined reference to `setsockopt' /usr/local/lib/libvga.so.1: undefined reference to `setgid' /usr/X11R6/lib/libX11.so.6: undefined reference to `unlink' /usr/local/lib/libesd.so.2: undefined reference to `strcasecmp' /usr/local/lib/libesd.so.2: undefined reference to `strtok' /usr/local/lib/libaudiofile.so.0: undefined reference to `fdopen' /usr/local/lib/libvga.so.1: undefined reference to `execv' /usr/local/lib/libvga.so.1: undefined reference to `setreuid' /usr/local/lib/libaudiofile.so.0: undefined reference to `__swbuf' /usr/local/lib/libvga.so.1: undefined reference to `getppid' /usr/X11R6/lib/libX11.so.6: undefined reference to `time' /usr/X11R6/lib/libX11.so.6: undefined reference to `poll' /usr/local/lib/libesd.so.2: undefined reference to `getitimer' /usr/local/lib/libvga.so.1: undefined reference to `seteuid' /usr/local/lib/libvga.so.1: undefined reference to `nice' /usr/X11R6/lib/libX11.so.6: undefined reference to `shutdown' /usr/local/lib/libesd.so.2: undefined reference to `getpwuid' /usr/local/lib/libaudiofile.so.0: undefined reference to `dup' /usr/lib/libncurses.so.5: undefined reference to `getcwd' /usr/local/lib/libesd.so.2: undefined reference to `gethostbyname' /usr/local/lib/libvga.so.1: undefined reference to `strsignal' /usr/X11R6/lib/libX11.so.6: undefined reference to `getpwnam' /usr/X11R6/lib/libX11.so.6: undefined reference to `getservbyname' /usr/local/lib/libvga.so.1: undefined reference to `fgetc' /usr/local/lib/libesd.so.2: undefined reference to `gethostname' /usr/local/lib/libesd.so.2: undefined reference to `strcspn' /usr/X11R6/lib/libX11.so.6: undefined reference to `setlocale' /usr/X11R6/lib/libX11.so.6: undefined reference to `getpeername' /usr/lib/libncurses.so.5: undefined reference to `strsep' /usr/local/lib/libesd.so.2: undefined reference to `fsync' /usr/local/lib/libvga.so.1: undefined reference to `valloc' /usr/lib/libncurses.so.5: undefined reference to `fputc' /usr/lib/libncurses.so.5: undefined reference to `cgetset' /usr/local/lib/libesd.so.2: undefined reference to `access' /usr/X11R6/lib/libX11.so.6: undefined reference to `rewind' /usr/lib/libncurses.so.5: undefined reference to `link' /usr/lib/libncurses.so.5: undefined reference to `cgetent' /usr/local/lib/libvga.so.1: undefined reference to `strspn' /usr/local/lib/libvga.so.1: undefined reference to `clock' /usr/lib/libm.so.2: undefined reference to `fputs' /usr/local/lib/libvga.so.1: undefined reference to `setegid' /usr/local/lib/libaa.so.1: undefined reference to `srandomdev' /usr/local/lib/libvga.so.1: undefined reference to `setuid' /usr/lib/libncurses.so.5: undefined reference to `mkdir' /usr/local/lib/libaudiofile.so.0: undefined reference to `frexp' /usr/X11R6/lib/libX11.so.6: undefined reference to `mbtowc' /usr/local/lib/libvga.so.1: undefined reference to `sysctlbyname' /usr/X11R6/lib/libX11.so.6: undefined reference to `strpbrk' /usr/X11R6/lib/libX11.so.6: undefined reference to `___tolower' /usr/local/lib/libvga.so.1: undefined reference to `raise' /usr/local/lib/libvga.so.1: undefined reference to `sigprocmask' /usr/X11R6/lib/libX11.so.6: undefined reference to `getsockname' gmake: *** [build] Error 1 -- Aaron From nobody Thu May 29 06:50:24 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Profiling with gcc Organization: ESC Date: 29 May 2003 06:50:24 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 34 kargl@troutmask.apl.washington.edu (Steven G. Kargl) writes: > Is SDL-1.1 a shared library? The missing routines appear > to be from libc_r_p, but libraries in the error messages > aren't being linked into build. Yes, I think it is shared. The Makefile runs 'sdl-config --libs' to get this part of it: -L/usr/local/lib -Wl,-rpath,/usr/local/lib -lSDL-1.1 -pthread I tried replacing that with 'sdl-config --static-libs', which returns a much longer list: -L/usr/local/lib -Wl,-rpath,/usr/local/lib -lSDL-1.1 -pthread -lm \ -L/usr/local/lib -lartsc -pthread -L/usr/local/lib -lesd \ -laudiofile -lm -L/usr/local/lib -lgnugetopt -L/usr/X11R6/lib \ -lX11 -lXext -lvga -laa -lusbhid That got rid of many of the undefined references, but not all of them. > Can you do > gcc -o build -g -pg -pthread build.o bstub.o engine.o cache1d.o \ > sdl_driver.o unix_compat.o a.o pragmas.o -L/usr/local/lib \ > -lSDL-1.1 -lvga -lesd -laudiofile -laa -lncurses -lX11 Yes, that worked! I had all those things when I tried static-libs, except for ncurses, and the order was different. Thanks very much! -- Aaron From nobody Thu May 29 21:46:27 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: looking for news server & client recommendation References: <60bd4c6b.0305262255.fa0d1fc@posting.google.com> Organization: ESC Date: 29 May 2003 21:46:27 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 Ben Crowell writes: > I just started using KNode as my client, and I really love it. It's > part of the standard KDE distribution. I was using emacs's gnus > package before, which lacked a lot of functionality, and was hard to > use. Heh, I can't argue with the "hard to use" part, but the reason Gnus is hard to use is that it has all the functionality you could ever possibly use, and then some. I've been using it for years, and I'm still finding new features. It's definitely not for newbies, though. I'd only recommend it to people with Emacs experience, preferably ones that can understand some very basic Lisp. -- Aaron From nobody Sun Jun 1 06:28:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: [Newbie] Local Subdomains References: Organization: ESC Date: 01 Jun 2003 06:28:12 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 44 Christopher Harrison writes: > This is how I see it: my network, connected to the Internet via a > cable modem, is behind (I think) a NAT router; so machines can't be > accessed directly (they don't have their own IP addresses, just > 192.168.0.x). After previous advice, I set my FreeBSD box's domain > name to an ISP qualified hostname, and its hostname to the > aforementioned "beastie" :) > This works fine, but it's all within the scope of our LAN > (i.e. local machines can access http://beastie/; but, externally, > http://beastie.[domain name]/ is unresolveable). When I started > reading about this DNS thing, everthing was with regard to the > big-bad outside world; and it also mentioned "people who don't know > what they're doing" ;) To make your virtual hosts (these aren't subdomains) reachable from the outside, they'll have to be resolved by the DNS on the outside that's responsible for the domain. So if your domain is "myhouse.com", the DNS(s) responsible for that domain will have to have a record for "beastie.myhouse.com" and any other virtual hosts you want to offer. They'll all need to point to your public IP address, which will need to be static. (It's possible to do this stuff with a dynamic address, but it adds another layer of headaches.) Do a `whois myhouse.com` to find out the DNS servers for your domain. You'll have to get the administrators for those machines to add your hostnames. Assuming all that can be done, you have two more steps. First, tell your cable modem router doohickey to forward all traffic on port 80 to your internal 192.168.0.x address, whatever it might be. Then config apache using the NameVirtualHost directive to let all your virtual hosts use that same IP. It doesn't sound like you have any need to run your own DNS, since other machines won't be doing lookups from it. Just put your hostnames in /etc/hosts with the machine's 192.168.0 address, to let apache match them up. -- Aaron From nobody Mon Jun 2 09:33:13 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie - FreeBSD5.0 and Win XP Pro References: <620f9833.0306020423.3b92fa2a@posting.google.com> Organization: ESC Date: 02 Jun 2003 09:33:13 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 36 bucon@eircom.net (Mark D) writes: > ifconfig output : > lo0: flags=8049 mtu 16384 > inet6 ::1 prefixlen 128 > inet6 fe80::1%lo0 prefixlen 64 scopeid 0x1 > inet 10.3.2.1 netmask 0xffff0000 This doesn't look right. lo0 is the loopback interface; it should always have the address 127.0.0.1. You should have another interface, called something0 (mine is ed0; it depends on the make of your network card) which has its address set to 10.3.2.1. In /etc/rc.conf, you should have a line that looks something like this: ifconfig_ed0="inet 10.3.2.1 netmask 255.255.255.0" The 'ed0' part will vary depending on your card. If this is set to lo0, that's causing your problem. lo0 will be setup automatically without any lines in /etc/rc.conf. If you don't know what to use in place of 'ed0', these commands might help find it: dmesg | grep -i ethernet dmesg | grep address If FreeBSD isn't recognizing your Ethernet card on bootup, you've got a separate problem. Find out what it is (by physically looking at it) and let us know the make and model. -- Aaron From nobody Mon Jun 2 11:34:00 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Newbie - FreeBSD5.0 and Win XP Pro Organization: ESC Date: 02 Jun 2003 11:33:58 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 talon@lpthe.jussieu.fr (Michel Talon) writes: > Besides that, when the FreeBSD box is configured correctly, it is > easier to use DHCP since Win boxes tend to configure the local > network via DHCP. So something like > ifconfig_ed0="DHCP" Ick. I don't get the fascination with DHCP on small networks lately. I can see using DHCP on a large network to avoid having to keep track of IP addresses for lots of different machines. (Although I'd still keep the addresses static, to aid in IP-level stuff like firewalling and packet tracing, and just use DHCP (or bootp) to centralize administration of addresses.) With two machines (or eight) this just isn't a concern. Assign an IP to the FreeBSD box in /etc/rc.conf, assign an IP to the Windows box in the network control panel, put their names and addresses in /etc/hosts, and be done with it. I just charged the owner of a small business for an hour of troubleshooting the other day because his ISP put in a DSL router and tried to change everything to use DHCP, and suddenly his PCs couldn't find some printers. Good for my finances, but he'd have been better off if they'd left his network alone. > On the WinXP side use the network wizard called "configuring a small > enterprise network". It will enable a dhcp server on the interface. > It likes to choose 192.168.0.1 for the Win box and something like > 192.168.0.137 for the other one. It will also allow to nat the > connexion to the web. Beware there is a firewall in WinXP which may > block packets, but the wizard should configure things correctly. Even more ick. Much better to use the FreeBSD box as the gateway, to protect the XP box behind it. At least that's the more common way to do it, so if you have trouble, you'll have an easier time getting help. -- Aaron From nobody Mon Jun 9 08:17:21 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: umask problem References: <44el23v3w0.fsf@be-well.ilk.org> Organization: ESC Date: 09 Jun 2003 08:17:21 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 12 Alex writes: > i know ,i set umask 037 for rwxr----- but it sets rw-r----- Right. By default, files are created with 0666 and directories with 0777, minus whatever bits are masked by umask. To make a new file executable, use chmod. -- Aaron From nobody Wed Jun 11 13:51:16 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: root login with ssh References: <20030611130249.34fb307c.einherjer01@hotmail-spammershaven.com> Organization: ESC Date: 11 Jun 2003 13:51:16 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 patpro writes: > nope, in fact I want the rsync to be launched by a script, on the > backup machine. I want to backup machine A to machine B and I want > to keep dates/ownerships/permissions for every file I rsync. Is there a reason why you can't run the backup script on A? I've got this situation, and I set "PermitRootLogin yes" on B and create a set of password-less keys allowing A to ssh to B automatically. Then I run my rsync process from A. I also ensure that B is fully firewalled from the net, even deleting its default gateway route and adding it manually just when I want to do cvsup or portupgrade, so that it's normally only reachable from my internal network. Another possibility is to run rsyncd on A, and use the rsync protocol instead of ssh to pull files to B. I'm not sure exactly how that's done, but it should be in the man page. -- Aaron From nobody Mon Jun 16 06:09:44 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Mustek 600 II CD scanner References: <3EEBDD7C.5070701@hotmail.com> Organization: ESC Date: 16 Jun 2003 06:09:44 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Christian Caron writes: > It is an old scanner I have. It is connected to an ISA SCSI > interface (DTC 3181E). This controller doesn't work under FreeBSD (or at least it wasn't supported when I tried it a year ago). It's garbage anyway, and on systems it does work with it brings the machine almost to a halt while it's scanning because it doesn't handle IRQs well or something. If you hook the scanner to a decent supported SCSI card, it should work fine. -- Aaron From nobody Tue Aug 26 10:05:31 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Dead serial ports Organization: ESC Date: Tue, 26 Aug 2003 10:05:30 -0500 Message-ID: <86y8xgo3wl.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 39 I've got a really odd serial port problem. I'm not even sure it's FreeBSD related, but I have to start somewhere... I had a system running 4.8-RELEASE on an old Pentium 200 system, dialing up ppp through an external modem on /dev/cuaa0 (COM1). All worked fine, until the hard drive had to be replaced. I thought I'd upgrade to 5.1 while I was at it. Everything worked fine, except ppp timed out waiting for a response from the modem. I tried talking directly to the modem, and while the TR light comes on, nothing I type does anything and the send/receive lights don't flash like they should. So I tried a different modem. Tried COM2. Different old motherboard (Cyrix 166). luck. Different serial cable. Tried going back to 4.6_2-RELEASE (what I happened to have on CD). Finally I even tried OpenBSD. Nothing worked; in every case I can't get any data through the serial ports, although the TR light on the modem lights up to show it's got a connection. The only way I can get it to work is to plug the modem into the serial port on a newer computer (Pentium 800), but I can't spare that for my dialup system right now, so that's not a long-term option. Anyone have any suggestions? As you can see, I've now replaced every component in the chain -- software, hardware, cables, etc -- and nothing helps. It sort of seems like an IRQ problem, but I don't know why that would have started happening all of a sudden, especially since this is a very bare-bones system -- just a video card and ethernet card; nothing to cause an IRQ conflict. Just had a thought: Is it possible for a faulty power supply to cause problems with the serial ports while everything else works? I think that's the only part I haven't swapped out. Thanks, -- Aaron From nobody Tue Aug 26 15:49:43 2003 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Dead serial ports References: <86y8xgo3wl.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Tue, 26 Aug 2003 15:49:42 -0500 Message-ID: <86u184m9eh.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "Jason Bourne" writes: > The serial ports make use of +/- 12 volts dc from the power > supply to generate the rs-232 signal. Check the 12vdc output > from the power supply; even if present it would be nice to view > it on an oscilloscope to look for excessive ripple. Thanks for the info. I don't have an oscilloscope, but I'll definitely try a different power supply and see if that helps. -- Aaron From nobody Sun Jan 4 20:25:31 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: NVIDIA driver problems References: <3ff20335$0$1182$636a55ce@news.free.fr> <10oIb.16342$aw2.9587535@newssrv26.news.prodigy.com> Organization: ESC Date: Sun, 04 Jan 2004 20:25:31 -0600 Message-ID: <86d69z6sis.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Scott Nightlinger writes: > Now, the situation that the nvidia driver seemingly produces the bug > in 4.x and not in 5.x is intriguing. Does anyone have the nvidia > drivers (not "nv", rather "nvidia") working in 4.x? They're working fine for me in 4.8-RELEASE. I don't remember all the details of what I had to tweak to get it working, but I reported my success in this group, so a Google Groups search should turn it up. -- Aaron From nobody Mon Jan 5 08:00:12 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: NVIDIA driver problems References: <3ff20335$0$1182$636a55ce@news.free.fr> <10oIb.16342$aw2.9587535@newssrv26.news.prodigy.com> <86d69z6sis.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Mon, 05 Jan 2004 08:00:12 -0600 Message-ID: <863cau5wcz.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 Scott Nightlinger writes: > dated 2003-01-28. It seems I may not have run the "make setup" > properly. I had read about it and attempted it by entering that > command but it did not do anything, just provided the next command > line. Maybe I should have investigated that situation more > thoroughly. Anyway, my system is working just fine with the 5.1 > -release. That's good to know. I've been thinking of upgrading, but wasn't certain the nvidia stuff would work. -- Aaron From nobody Thu Jan 8 08:41:08 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Mail Transfer Smarthost References: <074vc1-b3r.ln1@domain.invalid.org> Organization: ESC Date: Thu, 08 Jan 2004 08:41:08 -0600 Message-ID: <86u136sdtn.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 27 frank@SPAMLESSesperance-linux.co.uk (Frank Shute) writes: > If you've got Qmail already, why don't you put Qmail on your `dumb' > box too? The Qmail users mailing list will probably be able to help > you fix things up. There's a link to it from: > http://qmail.plig.org/top.html There's also this question in the FAQ: How do I forward unrecognized usernames to another host? With sendmail I had a LUSER_RELAY pointing at bigbang.af.mil. Answer: Put | forward "$LOCAL@bigbang.af.mil" into ~alias/.qmail-default. So he could setup his relay box with this and no local users, and insert any filters he wants to run before that 'forward' command in the pipeline. -- Aaron From nobody Thu Jan 29 07:33:41 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: which mail client can be integrate with PGP? References: <8i5ne1-v71.ln1@antioffline.de> Organization: ESC Date: Thu, 29 Jan 2004 07:33:40 -0600 Message-ID: <86hdyekhfv.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 "sam" writes: > I have a shell script send email notification perodically like this: > # mail -s "warning..." alert@company.com < /tmp/warning.txt > I want this the email msg is encrypted by PGP. gpg -a -e -o - -r whoever@wherever /tmp/warning.txt | mail -s \ 'warning...' alert@company.com -a does ASCII armoring, which you want because mail isn't going to turn it into an attachment for you. -e encrypts it. -o - sends it to standard output -r tells it which key in your keyring to encrypt for Then the result is piped to mail. -- Aaron From nobody Thu Feb 26 06:54:19 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: overly aggresive gnu-ghostscript package References: <7omqg1-b141.ln1@xor.obsecurity.org> Organization: ESC Date: Thu, 26 Feb 2004 06:54:19 -0600 Message-ID: <86brnmf1c4.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 39 hawk@slytherin.ds.psu.edu (Dr. Richard E. Hawkins) writes: > Stop in /usr/ports/print/ghostscript-gnu. > *** Error code 1 > > Stop in /usr/ports/print/cups-pstoraster. > *** Error code 1 Since the print/cups-pstoraster port is the one requiring print/ghostscript-gnu, I'd say that's where the problem is. From its Makefile (line broken to fit): RUN_DEPENDS= \ ${LOCALBASE}/share/ghostscript/${PORTVERSION}/lib/gs_init.ps:\ ${PORTSDIR}/print/ghostscript-gnu So if the file gs_init.ps doesn't exist, it'll try to install the ghostscript-gnu port. That should probably be changed to use the WITH_GHOSTSCRIPT_AFPL variable somehow, like putting these lines before that one: .if defined(WITH_GHOSTSCRIPT_AFPL) GHOSTSCRIPT_PORT=ghostscript-afpl .else GHOSTSCRIPT_PORT=ghostscript-gnu .fi And then replace 'ghostscript-gnu' in the RUN_DEPENDS line with 'GHOSTSCRIPT_PORT'. I'm not a port maintainer, though, so there's probably a more proper way to do that. For a quick fix, you could try just removing this dependency, or search your system for gs_init.ps and try to figure out why it's not where it's expected to be. Possibly cups-pstoraster won't compile against ghostscript-afpl if -afpl puts them in a different location than -gnu. -- Aaron From nobody Thu Feb 26 08:00:47 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Is sendmail really that insecure? References: Organization: ESC Date: Thu, 26 Feb 2004 08:00:45 -0600 Message-ID: <864qteey9e.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 55 "Peter Pippinger" writes: > i´m currently setting up an emailserver for some clients. It works > great very smoothly. But now i have heard that sendmail seems to be > a very insecure MTA? But if this is true, why does it come with > FreeBSD which is a very stable and secure system? Habit and knowledge, I suspect. It's been around forever, so people who have been doing this stuff a long time are familiar with it. Switching to a different MTA for the default install would be a big job, and would probably have to involve different people who knew the ins and outs of the new MTA intimately. > Are these all only rumours, or can i still sleep well at night if i > use sendmail as MTA? What bad things could happen by using sendmail? Well, I wouldn't be concerned about rumors, but you have to take security alerts seriously. If, like me, you're an administrator with clients who watch the security report lists regularly, one 'bad thing' would be having clients anxiously demanding that you upgrade their systems because they just got the latest sendmail exploit alert, even though it might be very low-risk. Of course, the worst bad thing would be someone breaking into one of your servers before you've dealt with a reported exploit. Sendmail is lots more secure than it used to be, but that doesn't necessarily mean it's great. I doubt even the strongest sendmail supporter would bet money that there won't be any serious exploits found in the next 12 months. By not using sendmail or bind, I'm able to ignore a lot of exploit reports. > What is the best MTA? I switched to qmail back in the olden days when the default sendmail install was still an open relay with lots of security risks, and you had to be fluent in sendmail's rules language to do anything unusual with sendmail.cf. I've never looked back. Qmail has had three minor revisions, from 1.00 to 1.03, in the time sendmail has had who knows how many major and minor revisions, despite the fact that sendmail was around long before qmail and should be more mature. Qmail plays nice with lots of other tools I like to use, and it's relatively easy to hack if you want to do something unusual, because it's broken up into lots of small tools. It's been around long enough and is used on enough busy servers now that its history of very few exploits can no longer be blamed on obscurity. The excellent step-by-step instructions at lifewithqmail.org make it a snap to set up, although it does take a couple hours. I'm not familiar enough with postfix or others to comment on them. -- Aaron From nobody Tue Mar 9 22:11:24 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Network Audio System Organization: ESC Date: Tue, 09 Mar 2004 22:11:24 -0600 Message-ID: <86n06pcpeb.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 18 I have two FreeBSD systems on my network. One (A) does not have a sound card. I'd like to use something like esound, nas, rplay, or wsoundserver to pipe the sounds from A over to the one with the sound card (B). I tried this with esound, but when I try to run esd on A, it complains that /dev/dsp does not exist. I suppose that's because without a sound card, the kernel never loads up the right module to create that device. Any suggestions how I can get around that? Also, any opinions on which network audio system is the best? None of them seem to have an abundance of documentation. I'll primarily be using it for the output from mplayer. Thanks, -- Aaron From nobody Fri Mar 19 00:57:47 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Error: Input: Resources temporarly unavailable References: Organization: ESC Date: Fri, 19 Mar 2004 00:57:46 -0600 Message-ID: <86fzc5coid.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 29 brent@dbprograms.com (Brent Bolin) writes: > Anybody seeing these error messages when running *vi* > > O/S 5.2-RELEASE-p2 I'm getting this occasionally lately when I run vi in an xterm on my 4.8-RELEASE system, with X 4.3.0. When it happens, vi won't run, and if I do an 'ls' on a directory with more than a few files in it, it quits partway through, but not always at the same point, and with no error message. I haven't updated the system lately, but various parts of X have been updated due to dependencies in the ports, including the XFree86-clients port (to 4.3.0_6), which contains xterm. At first I thought it showed up after running mplayer, but the last time it happened was soon after a reboot when I hadn't run mplayer yet, so now I know that's not it. I also started using NFS recently, but I hadn't mounted any NFS drives since this reboot, so unless the mere presence of nfsiod running in the background is causing it, I'd say that's not it. It's not a huge problem; I just exit the xterm and start a new one. It'd be nice to have it fixed, though. I suppose I should try some of the xterm alternatives and see if it shows up in them. -- Aaron From nobody Fri Mar 19 01:07:30 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: good, reliable window manager? References: <20040308202554.6facf85b.steveo@eircom.net> Organization: ESC Date: Fri, 19 Mar 2004 01:07:29 -0600 Message-ID: <86brmtco26.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Steve O'Hara-Smith writes: > flwm - it has failed to annoy me for some considerable time now. Heh. That's how I feel about window managers too: If they stay out of my way and don't annoy me, that's all I can ask for. I've been using Windowmaker for a couple years now, and haven't gotten annoyed enough to change. Probably partly because I've never bothered to learn to use most of the features. -- Aaron From nobody Fri Mar 19 01:24:41 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: burning hard drive to disk References: <405a2147$0$7056$be864849@news.hal-mli.net> <44ish1y5rw.fsf@be-well.ilk.org> Organization: ESC Date: Fri, 19 Mar 2004 01:24:41 -0600 Message-ID: <867jxhcn9i.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 31 Lowell Gilbert writes: > It's not quite that bad. dd can read from an offset, so you don't > need the free space to store the intermediate files. You could use > the same trick in reverse, to read a succession of image files to > their respective places on the disk. And you don't need the same > geometry to read it to, just anything as large or larger. Rather than run dd multiple times with different offsets, I'd just pipe it to split: dd if=/dev/mydrive | split -b 640m This will give you 640MB files named 'xaa', 'xab', xac', and so on. Then to write it back to another disk, just: cat x?? | dd of=/dev/newdrive If the disk isn't really full and you want to save some CDs, compress the data before splitting: dd if=/dev/mydrive | gzip -c | split -b 640m and to restore: cat x?? | gunzip -c | dd of=/dev/newdrive -- Aaron From nobody Fri Mar 19 09:47:07 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: burning hard drive to disk References: <405a2147$0$7056$be864849@news.hal-mli.net> <44ish1y5rw.fsf@be-well.ilk.org> <867jxhcn9i.fsf@cail.baugher.pike.il.us> <44k71h3qsn.fsf@be-well.ilk.org> Organization: ESC Date: Fri, 19 Mar 2004 09:47:07 -0600 Message-ID: <864qskc004.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Lowell Gilbert writes: > The whole point of the question was to avoid the extra space that > method requires to hold the output of split. Good point; my way does require that you have enough space to store all the images at once. The original poster didn't say whether that was a problem. With compression they could take a lot less space than the actual drive size, at least. -- Aaron From nobody Fri Mar 19 15:10:41 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD "newbie" question: How to change MTA in FreeBSD 5.2.1 References: <518783f3.0403191123.65158e77@posting.google.com> Organization: ESC Date: Fri, 19 Mar 2004 15:10:41 -0600 Message-ID: <86n06c8rvy.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 blixel@yahoo.com (David) writes: > My question is, how do I change the MTA? I run /stand/sysinstall, > then go to Configure, then Networking, then Mail, then I choose > Postfix. But the setting doesn't "take". Sendmail continues to run > (even after a reboot). I've never installed an MTA from the installer; I always install qmail from the ports. So I'm not sure how that's supposed to work. But I know you can stop sendmail from running on boot by setting "sendmail_enable" to "NONE" in /etc/rc.conf. That's one of the first things I do on all my installs. -- Aaron From nobody Mon Mar 22 10:20:52 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: maximum files per directory? References: Organization: ESC Date: Mon, 22 Mar 2004 10:20:51 -0600 Message-ID: <86n068onto.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 Richard Parker writes: > Don't do silly globs, use find and xargs, and if necessary split the > directories up, I've dealt with this by sorting the files into subdirectories according to name. In my case, all the files were named by a set of digits, like 1234567890123.txt. So I put them into subdirectories with the first level based on the first three digits and the second level based on the second set of three, like so: 123/456/1234567890123.txt That gave me up to 1 million potential different directories. If I needed to add a third level, it would have been easy enough. You could do the same thing with any sort of filenames, and hacking a program to get it to figure out the path should be fairly easy. In my case in Perl it was: $filename =~ s|^(...)(...)|$1/$2/$1$2|; Of course, you'd want to wrap some error checking around that to deal with filenames shorter than the system likes (maybe putting them in a special subdirectory) or with odd characters, but the basic concept is there. The best part about this, as opposed to say a system where you stick the files into subdirectories randomly and keep track with an index database of some sort, is that it's quite human-readable. A person can look at a filename and easily tell where it's stored if he needs to access it directly for some reason, like when restoring from backup. -- Aaron From nobody Mon Mar 22 14:49:46 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: 4.9 system locking up on xscreensaver Organization: ESC Date: Mon, 22 Mar 2004 14:49:45 -0600 Message-ID: <8665cwa9p2.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 32 I just upgraded from 4.8-RELEASE to 4.9-RELEASE yesterday. Since then, I've had two occasions where, when hitting a key to come out of xscreensaver (4.15), the screen seemed to freeze into a red/blue/black pattern, and I couldn't even kill X with Ctrl-Alt-Backspace. I logged in via SSH from another system, and found an unusually high load average, with X taking up 99% of the CPU. I killed xscreensaver -- no change on the screen. Killed X, dropping the load to normal -- still no change on the screen. Killed all the processes owned by my username -- still no change. The only way I could get the screen to change was with Ctrl-Alt-F1, which should take me to a console, but it only gave me a black screen. A ps listing from the SSH session showed gettys running normally. Eventually I gave up and rebooted. I've got nothing in dmesg or /var/log/messages to indicate what's going on. I never had this happen with 4.8. The only change I made in going to 4.9 was to add atapicam support and these two options to get mplayer to work better with my CPU: options CPU_ATHLON_SSE_HACK options CPU_ENABLE_SSE Does this problem sound familiar to anyone? As a test, I've limited xscreensaver to one routine that I've seen work fine. If it still locks up, I'll stop xscreensaver altogether for a while, and see what that does. I suspect it's not really the cause, though, or killing X would fix it. -- Aaron From nobody Wed Mar 24 18:37:09 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: 4.9 system locking up on xscreensaver References: <8665cwa9p2.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Wed, 24 Mar 2004 18:37:09 -0600 Message-ID: <86lllpyd6y.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 Martin Smith writes: > I have have had similar problems with xscreensaver in 4.9 stable > typically returning to the room after 10 mins to find xsceensaver > motionless and everything locked solid as you describe, I have just > stopped using it. It seems to be ok in 5.2 however. I told xscreensaver to only use a certain selected program (#20, which I think is jigsaw or something similar), and so far so good. If it doesn't lock up in a couple days, it should be okay. One thing I should have mentioned before is that I'm using the NVidia driver, and I didn't recompile that after rebuilding the kernel. I didn't know if that should be necessary, but that'll be my next move if I have more problems. I'm thinking that even if the problem originates with xscreensaver, it must involve something in the display driver or I'd be able to kill X and get back to a console. -- Aaron From nobody Thu Mar 25 11:18:48 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: 4.9 system locking up on xscreensaver References: <8665cwa9p2.fsf@cail.baugher.pike.il.us> <86lllpyd6y.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Thu, 25 Mar 2004 11:18:48 -0600 Message-ID: <863c7wvo93.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 Kris Kennaway writes: > Recompiling modules when you rebuild the kernel is absolutely > essential. Figures. I'll have to do that soon. It was kind of a pain getting the nvidia driver working the first time, so I didn't want to touch it unless I had to. I see the port has a much newer version, so I'll have to backup my current one and try that. So far, no more crashes, so apparently one of the xscreensaver routines was doing something that completely hosed the nvidia module. I commented out all the GL ones ago, so it wasn't one of those. Once I get the nvidia module recompiled, I'll let xscreensaver roll through all the possibilities again, and see how it goes. -- Aaron From nobody Thu Mar 25 11:33:46 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: small hdd install References: Organization: ESC Date: Thu, 25 Mar 2004 11:33:46 -0600 Message-ID: <86y8pou8zp.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 "Wayne" writes: > i have 2 1gig hdd's and wanna install freebsd. > > the machine i am using is a p2 450mhz, 128mb ram. > > if i do a auto assign of the disk label my swap is set to 240mb > which limits me greatly. > can anyone perhaps recommend the sizes i could use ? the machine > will be used as a small internet gateway/mailserver/webserver That sounds like plenty of swap for those services unless you plan to run X or some other resource hog on top of that. You might want to try it and see how much it uses; I've run those services and more on a lot less. On my home gateway, I've got a mail server (qmail), news server (leafnode), NFS server, and a handful of other common services running and it all adds up to about 72MB. Adding a webserver like apache shouldn't increase that greatly, unless it supports a busy site. It's typical to set your swap at double your physical RAM, and it sounds like that's about what it's doing. If you're actually using all that swap and more, you'd probably be better off adding actual RAM. The kind of add-ons that could push you into needing more memory, like a proxy server, SQL database, or spam checker, won't be all that happy if they're swapping full-time. One other thing: If the auto-assign is just setting up the first disk, you could manually add another swap partition on the second disk. When I've been in a similar situation, I've let the installer auto-create the usual partitions on the first disk, and then added two partitions on the second disk -- one for extra swap, and one containing the rest of the drive space to be mounted as /home. Then if space gets scarce on the first disk, I can always put things like /usr/ports or /usr/obj under /home and link to them. -- Aaron From nobody Thu Mar 25 17:52:43 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: new qt version (qt33) cannot compile on FreeBSD 4.8 - "ft2build.h hasn't been included yet"?? References: <1064tv32dv8u14c@corp.supernews.com> <10663eobbv1hl40@corp.supernews.com> Organization: ESC Date: Thu, 25 Mar 2004 17:52:43 -0600 Message-ID: <8665csscvo.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "Young Coot" writes: > After checking "freshports", it looks like a "libXft" commit (to > version 2.1.6) is the official fix for this - and, as luck would > have it, I did my most recent cvsup right before that. I had the exact same error when installing x11-toolkits/gtk20, as a dependency of www/firefox. Doing a portupgrade of libXft first solved it for me too. -- Aaron From nobody Fri Mar 26 05:30:56 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: auto start-up scripts References: Organization: ESC Date: Fri, 26 Mar 2004 05:30:55 -0600 Message-ID: <868yhnq1zk.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 sam writes: > can I move start-up scripts *.sh to /usr/local/etc/rc.d/ in freebsd > 4.9 if I want to make program start up automatically? I found > putting stuff in rc file is a bit messy. Yes, that's what that directory is for. Many programs will put a startup script there when you install them, but if they're well-behaved, they'll append ".sample" to the name of the script, so it won't run until you rename it so it ends in ".sh". For example, "mv mysql-server.sh.sample mysql-server.sh" will cause the mysql server to start on the next boot. -- Aaron From nobody Fri Mar 26 08:04:04 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Want to delete NTFS partition and use it for FreeBSD References: Organization: ESC Date: Fri, 26 Mar 2004 08:04:04 -0600 Message-ID: <86ad23ogbv.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 62 Boris writes: > I want to remove the Windows 2000 partition and convert it to UFS. I > tried this with sysinstall and fdisk which destroyed the partition > table. I managed to recover the partition table. FreeBSD does boot > again, but sysinstall claims the FreeBSD partition (ad0s3) is > unused. Is it safe to change the partition type to FreeBSD? I just did this yesterday with an MSDOS slice (type 6). (I'll try to refer to BIOS divisions as "slices" for clarity.) I wanted to change the second slice on the first IDE disk (ad0s2). First I used fdisk: fdisk -2 -u ad0 That lets you interactively change the slice info, just for the second slice (option -2). I left everything as-is except to change the type to 165 for FreeBSD. Then I wrote a disklabel to it with: disklabel -w ad0s2 Then I edited the disklabel: disklabel -e ad0s2 This brought up a disklabel in vi, with just the c partition created. I added a partition like so: # /dev/ad0s2: 8 partitions: # size offset fstype [fsize bsize bps/cpg] a: 61785990 0 4.2BSD 0 0 c: 61785990 0 unused 0 0 # "raw" part, I just matched the size to what the c partition had, since I just wanted one big partition, and set the fstype to 4.2BSD. After saving and exiting disklabel, I ran it once without options to make sure my changes took: disklabel ad0s2 Then I wrote a filesystem to the new partition: newfs ad0s2a You might want to add the -U option to that for soft updates. Then I added the partition to /etc/fstab: /dev/ad0s2a /st ufs rw 2 2 And mounted it: mount /st Done! -- Aaron From nobody Sat Mar 27 18:23:57 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: portupgrade eterm References: Organization: ESC Date: Sat, 27 Mar 2004 18:23:56 -0600 Message-ID: <86ekrdkeeb.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 "Understudy" writes: > I am trying to portupgrade eterm. I have tried several times. I have > tried a make deinstall && make reinstall. I have read through > UPDATING to see if anything was there about eterm. GIS results in > needed to make a change to the Makefile but that change was already > there when I examined it. looking to see if I have the ports > installed it is talking about I do: > imlib-1.9.14_2 > imlib2-1.1.0 I have those same versions installed, and I installed eterm without any problems a week or so ago. Have you cvsup'd your ports tree lately? My x11/eterm/Makefile requires Imlib2; it sounds like yours is trying to find imlib1 and failing. -- Aaron From nobody Sat Mar 27 18:28:21 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Want to delete NTFS partition and use it for FreeBSD References: <86ad23ogbv.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Sat, 27 Mar 2004 18:28:21 -0600 Message-ID: <86ad21ke6y.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 12 Boris writes: > Dos partitions are different. I discovered that fdisk can't delete > NTFS partitions. Really? Wow, how obnoxious is that. Oh well, maybe my instructions will help someone who needs to alter a DOS slice someday. -- Aaron From nobody Sun Mar 28 19:20:50 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Is there hope for my system after "portupgrade -a" has failed. References: Organization: ESC Date: Sun, 28 Mar 2004 19:20:49 -0600 Message-ID: <86isgoh2j2.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 Colin Andrew Percival writes: > This refers to the (short) time when portupgrade says "Updating > package database...", not the entire portupgrade process. Yeah, I kill portupgrade in the middle of downloads all the time. Usually it's because I've started installing something and then find I need to make it stop hammering my modem so I can get some work done online. -- Aaron From nobody Mon Mar 29 14:38:29 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Root filesystem full References: Organization: ESC Date: Mon, 29 Mar 2004 14:38:28 -0600 Message-ID: <86ekrbcrsr.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 26 Sean writes: > Filesystem Size Used Avail Capacity Mounted on > /dev/ad0s1a 248M 246M -18.4M 108% / > > Can anyone suggest how to begin troubleshooting this? cd / du -x -d 1 The '-x' says don't traverse other filesystems, and the '-d 1' says only go one level deep. That'll give you a list of all the directories in / that aren't on other filesystems, and how much space they're taking up. That should help you find what's filled it up. Use 'ls -l' to check for files, if the directory listing doesn't show anything large. If you see that a particular directory is using a lot, cd into it and run that same du command again to narrow it down further. I've got the same partition setup as you, and my / is only 72MB, so you've got something pretty big somewhere that should stand out. -- Aaron From nobody Mon Mar 29 20:02:00 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Root filesystem full References: <86ekrbcrsr.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Mon, 29 Mar 2004 20:02:00 -0600 Message-ID: <867jx3ay93.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 62 Sean writes: > Thanks for the reply. Before I had post my question I did a little > research and I learned about the command you had sugested. Below is > the output of that command: > 194M ./root There's your problem. Are you logging in and doing a lot of stuff as root? You don't want to do that. By the way, I hate the -h option for 'du'. It's a lot easier to spot the big directories if your listing looks like this: 72 ./dev 118 ./tmp 2 ./usr 2 ./var 2444 ./stand 1316 ./etc 2 ./cdrom 1 ./proc 2 ./dist 4124 ./bin 1054 ./boot 2 ./mnt 7806 ./modules 31424 ./root 12488 ./sbin 7628 ./modules.old 2 ./service That's a lot easier to read than the output of 'du -h' where you have to pick out the Gs and Ms from the Ks and Bs. Also, you can pipe it to a numerical sort: # cd /; \du -x -d 1 | sort -n 1 ./proc 2 ./cdrom 2 ./dist 2 ./service 2 ./usr 2 ./var 16 ./bak 72 ./dev 118 ./tmp 1054 ./boot 1316 ./etc 2444 ./stand 4124 ./bin 7628 ./modules.old 7806 ./modules 12488 ./sbin 31424 ./root Much nicer. -- Aaron From nobody Wed Mar 31 08:51:32 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Apache install References: Organization: ESC Date: Wed, 31 Mar 2004 08:51:32 -0600 Message-ID: <868yhh2hor.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 14 "SayWhat" writes: > Opensll comes with the base OS. Seems odd that > apache+mod_ssl-1.3.29+2.8.16_1 doesn't compile on a freshly updated > 4,9 REL box (2 in my case), don't ya think? Same thing on 5.1. The apache-modssl port only looks in /usr/local for openssl. Rather than try to convince it otherwise, I just installed the openssl port. -- Aaron From nobody Thu Apr 1 06:05:06 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: dual boot References: <84563411.0403311303.1e42e283@posting.google.com> Organization: ESC Date: Thu, 01 Apr 2004 06:05:06 -0600 Message-ID: <86n05vykct.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 "Alan McKay" writes: > i have to do this soon with win98 / freebsd > > is there howto guide anywhere? There's one in the Documentation Project: -- Aaron From nobody Wed Apr 7 19:24:31 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Net::DNS References: Organization: ESC Date: Wed, 07 Apr 2004 19:24:31 -0500 Message-ID: <861xmz8gg0.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 Nicolas Ecarnot writes: > In a perl software (OTRS), I need Net::DNS. > How can I install it, when /ur/ports/net/p5-Net-DNS is empty ? Many (all?) DNS related ports have moved to /usr/ports/dns. You can also use the CPAN module to install Perl modules: # perl -MCPAN -e "install Net::DNS" The first time you use CPAN, you'll have to answer some questions. You can accept most of the defaults, but you'll need to select a few servers from a list that best match your location. -- Aaron From nobody Thu Apr 8 07:27:43 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advanced mounting question References: <3f25c666.0404070644.454f8331@posting.google.com> <20040407213216.3d299f0b.steveo@eircom.net> Organization: ESC Date: Thu, 08 Apr 2004 07:27:43 -0500 Message-ID: <86d66i7iyo.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 11 Steve O'Hara-Smith writes: > NFS is your friend, mount it normally, export the bit you want > to localhost and mount it where you want it. NFS is no one's friend. -- Aaron From nobody Thu Apr 8 10:05:35 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advanced mounting question References: <3f25c666.0404070644.454f8331@posting.google.com> <20040407213216.3d299f0b.steveo@eircom.net> <86d66i7iyo.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Thu, 08 Apr 2004 10:05:35 -0500 Message-ID: <86lll65x34.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 fixx writes: > On Thu, 08 Apr 2004 07:27:43 -0500, Aaron Baugher wrote: >> NFS is no one's friend. > rotflmao, that is so true Seriously. I've tried NFS occasionally for about a decade now, on platforms from Next to HP to Linux to BSD with a few others thrown in, and every time I get annoyed and end up discarding it for a couple years. It's like getting back together with your ex-girlfriend -- it seems like a good idea when you're lonely and desperate, but before long all the reasons you broke up in the first place come rushing back. My latest excursion was between two FreeBSD machines in my home. Script on machine A reads images from my digital camera and puts them in a directory in a filesystem which is NFS mounted by machine B (which is not USB-capable, hence the run-around). So I plug the camera into A, run my script, and then look for the files on B. Not there. I log into A, check the directory, there they are. List the directory again on B, still nothing. Files that were already there are shown, but not the new ones. Wait a couple minutes, still nothing. This is taking 'asynchronous' way too far. I think the best network file system I've used was Apollo Domain's. It was so simple: by default, if you plugged a machine named "bannor" into the ring, every other machine could access its filesystem with "//bannor/pathname". I don't think it had the kind of security options NFS has, but for an internal network it was great. -- Aaron From nobody Thu Apr 8 17:38:27 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advanced mounting question References: <3f25c666.0404070644.454f8331@posting.google.com> <20040407213216.3d299f0b.steveo@eircom.net> <86d66i7iyo.fsf@cail.baugher.pike.il.us> <86lll65x34.fsf@cail.baugher.pike.il.us> <1081442854.609312@ente.ipberlin.com> Organization: ESC Date: Thu, 08 Apr 2004 17:38:26 -0500 Message-ID: <86y8p63xjx.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 jpd writes: > QNX can do such a thing, too. And it runs on faster hardware as > well. Are there any other network file systems for *BSD besides NFS, Samba, and Coda? It seems like Coda has been in experimental, "likely to break at any time" status for years. As for samba, it just doesn't seem right to have my BSD boxes communicate with a system that was designed for talking to Windows. -- Aaron From nobody Fri Apr 9 09:21:00 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advanced mounting question References: <3f25c666.0404070644.454f8331@posting.google.com> <20040407213216.3d299f0b.steveo@eircom.net> <86d66i7iyo.fsf@cail.baugher.pike.il.us> <86lll65x34.fsf@cail.baugher.pike.il.us> <1081442854.609312@ente.ipberlin.com> <86y8p63xjx.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Fri, 09 Apr 2004 09:21:00 -0500 Message-ID: <868yh52pwz.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Christopher Browne writes: > The CMU-sourced options include: > > - OpenAFS, a fork of AFS, which is available for a whole bunch of > Unixes; I may have to try this one. They don't have a FreeBSD version in their list of binaries, and it's not in the ports tree, so I'll have to build it and see how it goes. -- Aaron From nobody Fri Apr 9 09:22:45 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Advanced mounting question References: <3f25c666.0404070644.454f8331@posting.google.com> <20040407213216.3d299f0b.steveo@eircom.net> <86d66i7iyo.fsf@cail.baugher.pike.il.us> <86lll65x34.fsf@cail.baugher.pike.il.us> <20040408211316.7fc528ac.steveo@eircom.net> Organization: ESC Date: Fri, 09 Apr 2004 09:22:44 -0500 Message-ID: <864qrt2pu3.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 Steve O'Hara-Smith writes: > Most odd - I do this sort of thing a lot with NFS and it always > works fine for me. I suppose I could measure a delay but I would > have to try. Of course, it's quite possible that I did something wrong. But it's not like I'm a novice -- I've installed and used lots of software packages, and written a few -- yet somehow I always run into trouble with NFS. It's weird. -- Aaron From nobody Sun Apr 11 07:20:41 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: compiling / installing applications from sources other than ports References: Organization: ESC Date: Sun, 11 Apr 2004 07:20:41 -0500 Message-ID: <86lll2wvs6.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 27 "Graham Turner" writes: > how does one go about installing applications from sources other > than the ports collection ? It depends on the software. Once you unpack the archive, look for README* and INSTALL* files for instructions. In many cases, it's no more complicated than "./configure && make && make install", but always check. > qu2 - in installing / upgrading applications from sources other than > the ports collection is there not a danger that apps such as > portupgrade will cause issues as these will assume the 'best' source > to be the ports collection ? Yes, that's a possibility. When a port is checked for dependencies, that usually involves checking to see if a particular file exists on your system. If the version you install has that file, you should be fine. If it doesn't, and something else tries to install the port that you've already installed by hand, you'll need to edit the Makefile to tell it not to do that. And hope that the version you have will work with whatever other software depends on it. -- Aaron From nobody Mon Apr 12 08:11:42 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: ufs conversion References: Organization: ESC Date: Mon, 12 Apr 2004 08:11:42 -0500 Message-ID: <867jwltk6p.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 15 "T. Irmen" writes: > BTW: UFS is not bad, cause of journalling charcter and so, but how > gain directly access from windows, linux and so on... You can't. The only filesystem that all your operating systems will use is the msdos(VFAT) filesystem. Or put the UFS drive in a machine running FreeBSD, and mount it via Samba from your Windows machine. -- Aaron From nobody Wed Apr 14 08:55:44 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: which p2p music downloader do you recommend References: <1081873942.285920@news01.eclipse.net.uk> Organization: ESC Date: Wed, 14 Apr 2004 08:55:44 -0500 Message-ID: <867jwimzof.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 13 writes: > Much as I hate saying this, but I've settled on eMule for Windows > (which runs in vmware). Last time I've tried mldonkey (half a year > ago), it had an annoying memory leak, It still does. I just kill and restart it daily, when it's grown to about 128MB. It seems to do a pretty good job otherwise. -- Aaron From nobody Wed Apr 14 12:09:01 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: which p2p music downloader do you recommend References: <1081873942.285920@news01.eclipse.net.uk> <867jwimzof.fsf@cail.baugher.pike.il.us> Organization: ESC Date: Wed, 14 Apr 2004 12:09:01 -0500 Message-ID: <86ptaalc5u.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 16 writes: [about mldonkey] > You still lose your place in the queue for rare files when you kill > it. Ah, I didn't know that. Mine seems to spend most of its time searching, not downloading, but I don't know if it's queued up for any at that point. It's not unusual for it to show dozens of sources but zero active and an availability of zero on a file I've already partially gotten. -- Aaron From nobody Sat Apr 24 21:26:53 2004 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Creating CD from ccd/img/sub files Organization: ESC Date: Sat, 24 Apr 2004 21:26:52 -0500 Message-ID: <86smeskd1v.fsf@cail.baugher.pike.il.us> berkeley-unix) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 12 I have a set of files that I guess I'm supposed to burn to CD. There's a .sub, a .ccd, and a .img. The .ccd is 3KB, the .img is 547MB, and the .sub is 22MB. Can anyone tell me what to do with these, with either burncd or cdrecord, or some other tool? I tried just burning the .img file to disk, but that didn't give me something I could mount. Thanks, -- Aaron From nobody Wed Dec 27 10:56:12 2006 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: make buildworld w/limited disk space References: From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Wed, 27 Dec 2006 10:56:12 -0600 Message-ID: <86hcvhjn0z.fsf@cail.baugher.pike.il.us> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 31 "Max Files" writes: > Also, is there a way to use an nfs partition on another server to do > the buildworld? I have other FreeBSD-6.1 servers on this LAN with > bigger drives. If I mount the /usr partition on BIGSERVER-A can I > use those sources to buildworld and buildkernel on SMALLSERVER-B? I don't know any reason why not. I've mounted /usr/ports with NFS before, and was able to compile and install ports just fine. It'll be much slower, of course, with all the reading and writing over NFS, but it should work. > Where can I find docs on this process? There shouldn't be that much to it. If you already have all the source on smallserver, you could just mount an empty directory on bigserver as /usr/obj on this one: smallserver# mount -t nfs bigserver:/usr/local/emptyobj /usr/obj Then run your makes. That way all the building is being done on the remote drive, but when you do your installs, things that go other places in /usr will still go on the local drive. One thing: If you do any of the process in single-user mode, as is recommended, you'll have to manually mount the NFS drive even if you've added it to /etc/fstab. -- Aaron -- http://www.myspace.com/aaronbaugher "Take what you need and leave the rest." From nobody Fri Apr 27 10:20:53 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: command copy cp ; question References: <1oqt22gxk5wvv$.ojbjtx7lncai$.dlg@40tude.net> <86hcr3z9ud.fsf@mibsd.ka> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 27 Apr 2007 10:20:53 -0500 Message-ID: <864pn1kfmy.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 jokhva@bbn.cn (Denise H. G.) writes: > nawfer writes: >> and if I write so >> cp -r www/XXX www/YYY/ >> >> I have this >> www/YYY/XXX/ >> but I want so >> www/YYY/ >> I want take and copy only all the content inner at XXX and not also >> the container XXX. > I think, you'd better: > > cp www/XXX/* www/YYY/ That won't get any dot-files, though. Rsync will do it if you add a trailing slash to the source directory: rsync -r www/XXX/ www/YYY But rsync isn't part of the standard distribution, so a tar solution might be better: (cd www && tar -cf - XXX) | \ (cd www/YYY && tar --strip-components=1 -xpvf -) -- "Take what you need and leave the rest." From nobody Fri Apr 27 15:45:46 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: command copy cp ; question References: <1oqt22gxk5wvv$.ojbjtx7lncai$.dlg@40tude.net> <86hcr3z9ud.fsf@mibsd.ka> <864pn1kfmy.fsf@brinn.baugher.biz> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 27 Apr 2007 15:45:45 -0500 Message-ID: <86wszxh7gm.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 34 pakrat@localhost.pr.neotoma.org writes: > cp -pr www/XXX/. www/YYY Interesting; that does seem to work. I expected it to copy XXX *into* YYY, since XXX/. should be equivalent to XXX . So why does it copy the contents of XXX instead of XXX itself? By the way, `man cp` says -r is discouraged and doesn't copy symbolic links or special files. > Everyone's forgetting to preserve perms/timestamps > as well as the two directory entries in every directory. My tar example preserves permissions and timestamps. Why would you need to preserve . and .. ? Aren't those supplied automatically by the filesystem? > Some tars can be somewhat pathological if you are extracting > directories that are flagged as readonly. > gnutar was the first to fix this by noting these directories and > fixing afterwards. Most current implementations of tar appear to > have licked the problem but... True, but in a FreeBSD newsgroup, I figured other pathological tars weren't important. For that matter, not all cp commands have the same syntax, do they? -- "Take what you need and leave the rest." From nobody Sat Apr 28 22:03:44 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Hardware 3D on GeForce2 Integrated GPU From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Sat, 28 Apr 2007 22:03:42 -0500 Message-ID: <86zm4rg9v5.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 31 I'm trying to get the hardware accelerated 3D on my NVidia GeForce2 Integrated GPU to work, on FreeBSD 6.2-RELEASE with Xorg 7.2. I started with a clean install of the OS and sources, then added Xorg according to the instructions on the wiki: . I can use the 'nv' driver fine, but without any 3D acceleration, of course. When I use the 'nvidia' binary driver as recommended, the server always crashes with a fatal server error at the point where it tries to load the GLX extension. I tried the latest version, which apparently no longer supports this chip; then I tried 7184 and 9631 (from the ports) and 7185 (downloaded from NVidia). All crash in the same way. By the way, when I use the 'nv' driver, OpenGL programs complain, if that's relevant: Xlib: extension "GLX" missing on display ":0.0". So it seems GLX is an issue either way. My xorg.conf and logs from startx and the server are here: Any suggestions? I realize this is an old chipset -- the motherboard must be 6 years old -- but the docs for the binary driver list it as supported. Thanks, -- "Take what you need and leave the rest." From nobody Sun Apr 29 08:23:26 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Hardware 3D on GeForce2 Integrated GPU References: <86zm4rg9v5.fsf@brinn.baugher.biz> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Sun, 29 Apr 2007 08:23:26 -0500 Message-ID: <86r6q3fh69.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 41 Jason Bourne writes: > I believe the Xorg 7.x import is being worked on for > 7-Release. Until the bugs get worked out you may want to stick with > 6.9.0 for now with 6.2-Release. Just install the x11/xorg port and > then the x11/nvidia-driver-7184 port and be done with it. Unless, of > course, you want to upgrade to -current and help out debugging :-) Well, this quest started with me wanting to try out Beryl, and these instructions for that pointed me to the Modular Xorg install after installing 6.2-RELEASE. (Which says it's not that risky. It lies! :-)) The beryl port isn't in the standard ports tree, but only in the one with Xorg 7.2 that you get with git. I assumed from that that beryl only works with the newer Xorg. I suppose I can go back to 6.9.0, get the 3D working, and then see if I can get beryl working on top of that version. If I can't, at least I'll be able to use other OpenGL stuff at a frame rate higher than 1. (I wonder how much trouble it'll be to delete all the X ports and everything that depends on them and install it all again? More trouble than starting over with the 6.2 install disk, I suspect.) > Sometimes "bleeding edge" also means "not ready for prime time". It > does look like you are real close to having it run (from your logs), > but if you aren't in a position to nail down why it segfaults it > might be better to drop back to something known to "just > work". Depends on where you want to spend your time. Well, I'd be willing to try to figure it out, but it's probably out of my league. Thanks! -- "Take what you need and leave the rest." From nobody Tue May 1 13:19:32 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Hardware 3D on GeForce2 Integrated GPU References: <86zm4rg9v5.fsf@brinn.baugher.biz> <86r6q3fh69.fsf@brinn.baugher.biz> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Tue, 01 May 2007 13:19:31 -0500 Message-ID: <86y7k8v230.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 25 Jason Bourne writes: > Don't bother - I don't believe Beryl will work with 6.9.0. And from > the couple of times I played with it on a Linux box it crashed. I > don't think it's completely ready yet either - the Novell OpenSuse > and Redhat people may be jumping the gun. I have seen where people > have reported using it successfully, but I've not had the luck. So I > wait for the code to get better. Give it time and it will straighten > out 'cause I know people are working on it. That's pretty much the same conclusion I came to. I've been using windowmaker for several years, so I think I can live with it a while longer. It does look like a busy project. I just started over from scratch -- 6.2-RELEASE, Xorg 6.9.0 from the ports on top of that, the 7184 nvidia driver from ports on that, and now 3D works fine. Glxgears reports over 10 times the framerate I was getting with the 'nv' driver, so I'm happy. On to installing EDuke32! Thanks, -- "Take what you need and leave the rest." From nobody Wed May 2 15:27:04 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: command copy cp ; question References: <1oqt22gxk5wvv$.ojbjtx7lncai$.dlg@40tude.net> <86wszxh7gm.fsf@brinn.baugher.biz> <44fy6g14qb.fsf@Lowell-Desk.lan> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Wed, 02 May 2007 15:27:04 -0500 Message-ID: <86slaft1if.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 29 Lowell Gilbert writes: > bv@wjv.com (Bill Vermillion) writes: > >> In article , Bill Vermillion wrote: >> >> ... >> >>>And all subdirectories have a . and .. entry. >>> >>>And you do want to preserve those persmissions when you are >>>restoring. >> >> And all the stuff I wrote got repeated about 5 times. >> I accidentally hit a wrong key in vi and duped the pgh's I had >> written and I was sure that I had deleted them all. >> >> My apologies. Up is a direction in which I don't screw to often :-) > > Oh, well. I thought you had just made the most amusing recursion joke > I had seen in years. Heh, that's exactly what I thought it was. -- "Take what you need and leave the rest." From nobody Fri May 4 09:19:21 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Any Fix For Gettext/Updating? References: <463a6784@news.broadpark.no> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 04 May 2007 09:19:21 -0500 Message-ID: <86k5vor7rq.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 Torfinn Ingolfsen writes: > 3. when you install FreeBSD (any version), do not (I said: do NOT) > install the ports tree that is on the CD. > 3. Download (or use cvsup) a fresh ports tree from the net Is there any reason it would be harmful to install the ports from the CD and then immediately update them? That's what I've always done, assuming cvsup would be faster that way, since there would be many files that wouldn't have changed since the ports archive on the CD was created. In my dialup days, a fresh ports tree could take a day to cvsup; but even now at 768Kb, it takes a while. -- "Take what you need and leave the rest." From nobody Thu May 17 17:07:34 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Partitioning Help References: From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Thu, 17 May 2007 17:07:34 -0500 Message-ID: <86odkj14ux.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 24 Earl Larsen writes: > I want to change my Win box to a FreeBSD box. I have 3 partitions on > the hdd. > Part 1 Win installation ect. > Part 2 swap > Part 3 media, downloaded ect files > I want to change part 1 and 2 to FreeBSD and keep the part 3 > alone. I want to access part 3 with FreeBSD. All of the partitions > are formated in NTFS. Delete the first two slices, then create one new FreeBSD slice in that space. Divide it up with disklabel and install FreeBSD in it, then once your system is up and running, mount the second (formerly third) slice as NTFS wherever you want it. -- "Take what you need and leave the rest." From nobody Fri May 18 08:33:12 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Partitioning Help References: <86odkj14ux.fsf@brinn.baugher.biz> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 18 May 2007 08:33:12 -0500 Message-ID: <86646q1ckn.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 20 Earl Larsen writes: > hmmm so If I mount it in fstab. I can mount it as auto or ntfs, and > it will work? Would their be any limitations? I don't know if "auto" will work, but "ntfs" should. See 'man mount_ntfs' for the limitations. The last time I had an NTFS partition a few years ago, you could only mount them read-only, but it appears that writing is now supported, with a few caveats. # mkdir /porn # echo /dev/ad0s2 /porn ntfs rw 0 0 >>/etc/fstab # mount /porn -- "Take what you need and leave the rest." From nobody Tue Jun 19 22:04:12 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Xorg 7.2 Listening On Port 6000 References: From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Tue, 19 Jun 2007 22:04:12 -0500 Message-ID: <86wsxz498z.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 19 per@hedeland.org (Per Hedeland) writes: > I think this has pretty much been obsoleted by ssh - why use the plain > X11 TCP transport that is either simple to use and totally insecure, or > semi-complex to use and semi-secure, when ssh's X11 forwarding gives you > "complete" security with zero effort on your part? I still use the old xhost method on my private network behind a firewall, because my machines are old enough that there's a noticeable slowdown if I run all my remote X apps through the encryption/decryption of ssh. Over a public connection, though, I generally let ssh do it. -- "Take what you need and leave the rest." From nobody Thu Jun 21 08:20:19 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: upgrading xorg? References: <466b00b2$0$8681$e4fe514c@dreader18.news.xs4all.nl> <466bf043$0$19720$e4fe514c@dreader26.news.xs4all.nl> <1181495542.40424@news.queue.to> <466c3c8e$0$30303$e4fe514c@dreader19.news.xs4all.nl> <466c58fc$0$18456$e4fe514c@dreader16.news.xs4all.nl> <466cef00$1@news.broadpark.no> <4679ab27$0$20288$9b4e6d93@newsspool3.arcor-online.net> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Thu, 21 Jun 2007 08:20:19 -0500 Message-ID: <86hcp130mk.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 35 Christoph Weber-Fahr writes: > No. It is completely unacceptable. > > In the past the ports system did not force you to upgrade > your X to the latest bleeding edge version whenever you > touched anything. The one you had usually did just fine. I have two ports trees on my system now: one normal one in /usr/ports.new, and one in /usr/ports.old that I rolled back with cvsup to the day before the xorg7.2 merge. I link /usr/ports to one or the other as needed. It's not that I'm unwilling to do the upgrade -- this is my desktop machine, so downtime isn't a big issue -- but when I tried to upgrade a couple weeks before the merge, I had problems with my old integrated NVidia GPU not being able to handle the 3D stuff, causing X to crash when it loaded GLX. It works fine under 6.9, and I only expect to use this machine for a few more months anyway, so I figured I'd just wait until I replaced it to go to 7.2. I was disappointed to find that I could no longer install or upgrade *any* ports requiring X at all. Didn't you used to be able to install X-dependent ports whether you used Xorg or Xfree86? I was surprised something like that wasn't done, or moving 6.9 to x11/xorg6.9, like is done with many other ports that people don't all want the latest version of. I'm certainly not a ports developer, though, so I have to assume that wasn't possible for some reason. -- "Take what you need and leave the rest." From nobody Fri Jul 27 21:49:02 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Nub question: FreeBSD is graphic or text based? References: <1Swni.130014$NV3.124169@pd7urf2no> <46aa02f1$0$28159$9a6e19ea@unlimited.newshosting.com> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 27 Jul 2007 21:48:59 -0500 Message-ID: <86wswlb5us.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 17 bv@wjv.com (Bill Vermillion) writes: > I read news in text mode, still using trn - as it works well for > me. And I have a news-server on my own machine for blazing speed > :-) Back when I did a lot of my Unix work logged in from a Commodore 128 running Craig Bruce's 80-column ACEterm, I did everything from within Emacs -- mail, news, shells, irc, everything. It was a bit limiting, but ever so cool. -- "Take what you need and leave the rest." From nobody Thu Aug 9 08:48:17 2007 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Nub question: FreeBSD is graphic or text based? References: <1Swni.130014$NV3.124169@pd7urf2no> <46aa02f1$0$28159$9a6e19ea@unlimited.newshosting.com> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Thu, 09 Aug 2007 08:48:17 -0500 Message-ID: <86lkckzun2.fsf@brinn.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 bv@wjv.com (Bill Vermillion) writes: > In the UUCP days I had two Telebit Trailblazers - both running near > maximum often - as I was feeding about 20 or 30 sites - some from > out of state. That's exactly how we got our news feed at the ISP I worked for in the mid-90s: one Telebit at our place, the other at a company we worked for on the side that had a news feed over their T1. Running UUCP at 9600bps, with that special mode they had when talking to each other than pushed them up around 19.2, they seemed to be loaded around the clock. When that couldn't keep up anymore, we got a satellite feed (can't remember the name of the company), but that was horrible on reliability, so we kept using UUCP for a while to fill in what the dish missed. -- From nobody Fri Jan 25 06:32:27 2008 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Graphics hardware compatibility References: <7eiiqyhtrgb9.dlg@mid.crommatograph.info> From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Fri, 25 Jan 2008 06:32:27 -0600 Message-ID: <86sl0m2ihw.fsf@bannor.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 22 Oliver Cromm writes: > I find the compatibility issues quite confusing. > > The new machine I have set my eyes upon leaves me two choices for the > graphics: > > - ATI Radeon X1250 256MB Shared Graphics featuring ATI Avivo Technology > - nVidia GeForce 8500 GT 256MB PCI-Express *Fanless* I'm using this nVidia card successfully with FreeBSD 6.x-i386, xorg 7.3, nvidia-driver-100.14.11 (the version that happened to be current in ports when I installed it), and compiz as my window manager, on a dual-core x64 system. I've had one X lockup in about 3-4 months since I got the system; no idea what caused it. -- "Take what you need and leave the rest." From nobody Mon Feb 4 18:51:20 2008 Newsgroups: comp.unix.bsd.freebsd.misc Subject: X lockups From: Aaron Baugher Organization: Baugher Salvage and Really Wild Stuff Corporation Date: Mon, 04 Feb 2008 18:51:20 -0600 Message-ID: <86ir14mdiv.fsf@bannor.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 28 Does anyone know what these error messages during an X lockup might mean, or where I should start looking? kernel: NVRM: Xid (0002:00): 8, Channel 00000001 kernel: NVRM: Xid (0002:00): 30, L1 -> L0 kernel: NVRM: Xid (0002:00): 13, 0001 00000000 0000502d 00000860 \ 00000000 00000100 This system was fine for the first few months, but in the last couple weeks, X has locked up on me three times. I haven't updated any software lately, so I tend to suspect the video card. If I use Ctrl-Alt-Backspace, X eventually dies (or maybe it dies on its own), but it takes a few minutes, and if I try to restart X, it's extremely slow and seems to freeze again. Rebooting brings it back to normal. I've been using compiz, so I've switched to windowmaker for now to see if that makes any difference. Other details: FreeBSD 6.1-RELEASE-p20 xorg-7.3_1 nvidia-driver-100.14.11 GeForce 8500GT Thanks, -- From nobody Sun Dec 11 07:28:53 2011 Newsgroups: comp.unix.bsd.freebsd.misc Subject: How Is FreeBSD on the Desktop These Days? Organization: Baugher Consulting Date: Sun, 11 Dec 2011 07:11:47 -0600 Message-ID: <86y5ujs164.fsf@baugher.biz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Lines: 68 I'm looking for opinions on using FreeBSD (especially amd64) on a desktop workstation, specifically for web programming, meaning that I need a browser that works correctly, including Flash (fortunately Flash seems on the way out, but it's not there yet). I'd also like to be able to do some video editing, and I like compiz (with some 3D effects) as a window manager. So in general, how is FreeBSD for a desktop these days? I'm fine with the old text installer; I'm only concerned with desktop functionality and apps on the running system. Here's the whole story: I was a pure FreeBSD user for several years, after switching from Linux. On the server, FreeBSD simply can't be beat, and I still use it there. But a couple years ago, I was getting frustrated with the fact that Adobe's Flash plugin for FreeBSD was always a version or two behind the Linux one, and especially on 64-bit it was unreliable. (Stupid closed-source Adobe's fault, not FreeBSD's.) There were some other annoyances: Adobe Air seemed to be becoming popular, and it didn't work well either. OpenOffice had an issue with the spellchecker not working, though I can't remember if that was FreeBSD specific. Seems like I also ran into trouble getting an onboard network card to work. There were also conversations on this newsgroup at the time about how the ports tree was getting awfully large and developing dependency issues like the Linuxes tend to have in their binary package managers. I got to a point where I felt like I needed to wipe my system and start over anyway. So I switched to Ubuntu (Xubuntu, actually), and I was pretty happy with it for a while. I missed the ports tree and some favorite utilities (why no cpdup on Linux?), but Flash was pretty solid, and I was able to get my work done. I've never come to *like* Linux as much as FreeBSD, though. It tries a little too hard to abstract things away from me, and nothing aggravates me more than not being able to do something like configure a network card because a problem with the window manager has wiped out the icon for it. I feel like there's more installed and running on the system that I'm not aware of, and it's not as good at logging things I need to know. I'm just much happier with FreeBSD's way of doing things. On top of that, Ubuntu has developed its own issues, at least on my system. If you don't stay right up on the latest version, you tend to get old versions of third-party software too. On FreeBSD, if a port won't compile unless you upgrade the OS, you'll find that out when you install it. But on Ubuntu, you just get whatever old version the repository last offered for your OS version. I don't blame them for not keeping up on older versions of everything, but when I'm only one upgrade behind, it doesn't seem like it should make as big a difference as it does. And now Flash has started working poorly. About 1/3 of the time I load a YouTube page, the video is gray (Flash fails completely) or the audio plays to a black background. Other sites with Flash cause the plugin to die or lock the browser up for a while. In addition to that, there's some weirdness with the menus in XFCE4 and compiz, where I have to mouse over a menu item twice to get its submenu to show up. Tooltips in all apps are the same way, and a bug report on that has been languishing for a year. And to top things off, now I'm getting spontaneous reboots, with nothing at all reported in any log. They're happening when the screensaver wakes up, and I've tried having it do nothing but blank the screen, or don't blank the screen and just play a (non-3D) screensaver, to no avail, so I'm pretty sure it's not a hardware thing. So I'd really, really, really like to switch back. Any encouragement or warnings? Thanks, -- From nobody Fri Sep 14 14:08:07 2012 Newsgroups: comp.unix.bsd.freebsd.misc Subject: LG DVD Burner Issues Organization: Baugher Consulting Date: Fri, 14 Sep 2012 14:08:06 -0500 Message-ID: <86mx0s4e3t.fsf@bannor.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain Lines: 48 I recently switched my system from Ubuntu to FreeBSD 9.0-amd64. It has an LG GH22NP20 DVD burner on the ATAPI bus, reporting firmware 2.00, which appears to be the latest. It worked fine under Ubuntu. Under FreeBSD, it reads discs fine that I burned under Ubuntu, but it's unable to read discs that it burns itself. I've burned disks with growisofs, and also with k3b (which I think uses growisofs). Neither gives me a disc the drive can then read. growisofs fails about half the time with output like this: Executing 'builtin_dd if=backup.iso of=/dev/pass2 obs=32k seek=0' /dev/pass2: "Current Write Speed" is 4.1x1352KBps. 0/4695836672 ( 0.0%) @0x, remaining ??:?? RBU 100.0% UBU 0.0% 0/4695836672 ( 0.0%) @0x, remaining ??:?? RBU 100.0% UBU 0.0% :-[ WRITE@LBA=0h failed with SK=3h/ASC=0Ch/ACQ=80h]: Input/output error :-( write failed: Input/output error But if I try it again, it may go ahead and work. When it works, it finishes and closes the disc without any complaints, but then I can't mount the disc. I can mount the ISO the disc was created from. I've tried several different blank DVDs. I get a lot of stuff like this in the kernel log: (cd0:ata0:0:0:0): READ(10). CDB: 28 0 0 0 0 10 0 0 1 0 (cd0:ata0:0:0:0): CAM status: SCSI Status Error (cd0:ata0:0:0:0): SCSI status: Check Condition (cd0:ata0:0:0:0): SCSI sense: MEDIUM ERROR asc:10,0 (ID CRC or ECC error) (cd0:ata0:0:0:0): cddone: got error 0x5 back g_vfs_done():cd0[READ(offset=32768, length=2048)]error = 5 Once I've burned the disc (and I can see the track on the disc) and try to mount it, I get that same "error = 5" line in the kernel log, and mount says: mount_cd9660: /dev/cd0: Input/output error I don't have another DVD drive to try these discs in, but I'm going to borrow one this weekend. Any suggestions what else I can try in the meantime, or other commands I can run to get more information? If this drive just isn't very compatible with FreeBSD, any suggestions on what DVD burner would be? An external one would be nice, but whatever works. Thanks, -- From nobody Thu Sep 27 09:47:45 2012 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: LG DVD Burner Issues Organization: Baugher Consulting References: <86mx0s4e3t.fsf@bannor.baugher.biz> Date: Thu, 27 Sep 2012 09:47:44 -0500 Message-ID: <86a9wbmsj3.fsf@bannor.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain Lines: 18 > I recently switched my system from Ubuntu to FreeBSD 9.0-amd64. It has > an LG GH22NP20 DVD burner on the ATAPI bus, reporting firmware 2.00, > which appears to be the latest. It worked fine under Ubuntu. Under > FreeBSD, it reads discs fine that I burned under Ubuntu, but it's unable > to read discs that it burns itself. Thought I should follow up on this: I opened up the burner, blew out the dust, cleaned some greasy dust off the rails, and cleaned the lens. That seems to have fixed it. I've burned and read a dozen disks since without any trouble, except for one coaster that had an obvious scratch on it. I'm not sure why it became an issue when I switched operating systems; maybe just coincidence. -- From nobody Tue Feb 2 17:03:26 2016 Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Binutils Organization: Baugher Consulting References: Date: Tue, 02 Feb 2016 17:03:25 -0600 Message-ID: <8660y6u342.fsf@bannor.baugher.biz> MIME-Version: 1.0 Content-Type: text/plain Lines: 17 Louis Epstein writes: > Both Portmaster and Pkg want me to "upgrade" binutils-2.26 to 2.25.1. > > Counterintuitive,what's the explanation? According to the commit history found here: Revert update to binutils 2.25.1 and bump portepoch to reflect that revert binutils has build issues on 10.x i386 and is full of regressions on 9.x: unexpected build failures on nss and others --