Re: version control systems
On Sat, Jan 09, 1999 at 03:59:07PM +1100, Hamish Moffatt wrote: On Fri, Jan 08, 1999 at 09:22:26AM -0800, Mike Touloumtzis wrote: I use Perforce at work and have been happy both with its quality and with the responsiveness of Perforce support. Their software has run on Linux for quite a while now, which is a nice positive. The Windows client looks a bit average. The mechanism doesn't APPEAR to cope with a development environment like Delphi, where you move between files all the time. It wants to load your editor for a file, then close it and open it again on the next. The GUI client is for wimps ;-) Most people here don't need it and don't feel a strong need for it; the command line client is pretty full-featured. But it may not be what you need. We chose it because it handles branching extremely well; if you don't expect to be in a situation in which the code will frequently be branched into sub-projects/working snapshots/etc., its advantages over CVS start to decline. miket
Re: Why not as a newsgroup?
On Sat, Jan 09, 1999 at 08:12:40PM +0100, Thomas Adams wrote: On Fri, Jan 08, 1999 at 01:09:59AM -0800, Martin Waller wrote: I for one would no longer to be able to get the list. I access the internet through a firewall. The policy is restricted viewing only of some newsnet news groups. There is no way to either connect to a different news server nor post to usenet. So the newsgroup way would be impossible for me to access. O, pity you... Can't afford a private ISP? Come on, this is the most ridiculous argument I've ever read. If your employer's net access doesn't work properly, get a real ISP account for home. Whether or not you choose to acknowledge it, the firewall issue is a real one for many people (myself included). miket
Re: fvwm95 menus
If you want to completely override the default values, just copy the default configuration file in full to the user's directory (as .fvwm95rc or whatever it uses) and edit it to remove the Read statements (for the hooks) and to change the menus to just the apps you want. miket On Sat, Jan 09, 1999 at 04:15:38PM -0500, Matt Kopishke wrote: What's the simplest way to completly over ride the menus the are used by default under fvwm95, for certen users (probly using ~/.fvwm95 and hook files?). I need a simple menu (to apps, Netscape Wordperfect) for our xterinals, the default debian-menu is confusing, and give them access to too many Apps. Thanks, -Matt-
Re: Kernel 2.2.
On Thu, Jan 07, 1999 at 01:32:27PM +0100, Thomas Janke wrote: The new LINUX-Kernel will be available soon. Is Debian ready for the new kernel? Will Slink be? I'm running 2.2.0pre4 on Slink right now with no apparent problems, so even though Slink won't likely ship with 2.2, you can upgrade easily. I compiled and installed this kernel using Debian kernel-package. Check out the kernel's CHANGES file to see other upgrades needed to make the jump to 2.1/2.2: http://www.linuxhq.com/change21.html Make sure you read this before you report any problems. As far as I can tell, the only package in Slink that doesn't make the cut version-wise is sysutils; 'procinfo' segfaults on the newer kernels. It's something most people can live without, although it's nice. miket
Re: Vanishing commands.
I think hwclock supplanted clock. miket On Thu, Jan 07, 1999 at 04:28:42PM +, Nidge Jones wrote: Since upgrading to 2.0, I have lost some commands :-( /sbin/clock has vanished for starters ! Where did this go ! Have I missed a package out somehwere - if so which ! ... (snip) The /sbin/clock command is a must though ! -- Nidge Jones
Re: A few (simple?) questions...
On Thu, Jan 07, 1999 at 01:03:12PM -0500, David Kennedy wrote: Hi, I was able to get Debian installed and everything I need working but I have a few questions: 1) How do you change the colours of xterm? Right now, when I click on it I get a black on white window. I can type xterm -bg black -fg white to get a new xterm window with the colours I want. Changing the menu to insert the -bg and -fg switches is window manager specific. I use fvwm2, and created a ~/.fvwm2/main-menu-pre.hook with this in it: # Same as main-menu.hook, but entries are added at the # very beginning of the menu. # for example: + rxvt Exec rxvt -ls -sl 256 -bg linen -geometry 80x56 + rxvt (big) Exec rxvt -ls -sl 256 -fn 10x20 -bg linen -geometry 80x56 + rxvt (black) Exec rxvt -ls -sl 256 -bg black -fg white -geometry 80x56 + rxvt (black, big) Exec rxvt -ls -sl 256 -fn 10x20 -bg black -fg white -geometry 80x56 + Emacs Exec emacs -g 80x70 + Emacs (I18N) Exec emacs -g 80x70 -fn fontset-standard # and the following generates a horizontal line: + Nop 2) What is dwww? I noticed in the menu directory where all the names of entries into the debian menu are that lots of choices aren't being displayed because they require 'dwww'. 3) How do I set a screensaver to activate? The menu lists many but I don't know how to set them. Also, I downloaded the opengl(?) screensavers and I can run them but they aren't listed under the screensaver section. http://www.debian.org/Lists-Archives/debian-user-9812/msg01163.html miket
Re: version control systems
On Fri, Jan 08, 1999 at 10:38:01PM +1100, Hamish Moffatt wrote: Summary: can anyone recommend a good multi-user version control system with Windows and Unix support, that runs in an acceptable time on a WAN, which supports binary files? cvs almost fits the bill, but I need to store Delphi source files in it, half of which are binary. I would also like a Windows front end for some of our users, but that's not essential. I use ClearCase at work. I like the feature set, but $4000 per single user license (US), with no Linux support? No thanks! Perforce (perforce.com) is $600 per user and they have Linux client and SERVER support. That's a nice touch. I use Perforce at work and have been happy both with its quality and with the responsiveness of Perforce support. Their software has run on Linux for quite a while now, which is a nice positive. miket
Re: Background(color)
On Wed, Jan 06, 1999 at 09:16:05AM +0100, CUNO wrote: Hello, Can I change the background(color) of the xdm login screen. Thanx Cuno Look at the files in /etc/X11/xdm/, especially 'Xreset' and 'Xsetup'. You can add e.g. xsetroot -solid aquamarine to one or both. Try experimenting; I did this a while ago on another machine and I don't remember the exact setup I ended up with. miket
Re: Winmodem?
On Fri, Jan 01, 1999 at 10:39:28AM -, eferen1 wrote: Hi. I had that problem too. If it's a winmodem, forget it! Get yourself a real modem. Debian people won't write a driver for it. s/Debian/kernel/ s/won't/can't/ Trust me, if hardware specifications for this WinModem were freely available, some signal processing sicko somewhere would be working on a driver for it. MS has a closed door policy on their code, and probably locks their developers in it too. As evil as MS are, they are not responsible for all of the world's problems. Blame US Robotics for this one; they have resisted repeated pleas for open WinModem specifications. Send them a (polite) note if you have one of their WinModems and don't like their policy. miket
Re: Cleaning my /usr/lost+found
On Thu, Dec 17, 1998 at 11:29:38AM -0500, [EMAIL PROTECTED] wrote: The funny (if that's the word) thing is, I once had a few errors on this drive but WD's Tech Support walked me through test/diagnostics. The number of errors then was too low for a warrenty repair. Thoug they said if I had anymore erorrs it'd be covered under warranty. Looks like I get to test WD's support. Joy. Push them to the wall on this one. I experienced exactly the same problem with a 2GB WD Caviar drive. It's a firmware bug; there's a reference to it hidden on the WD homepage. miket
Re: Cleaning my /usr/lost+found
On Fri, Dec 18, 1998 at 03:56:37PM +1100, Hamish Moffatt wrote: On Thu, Dec 17, 1998 at 04:53:33PM +0100, E.L. Meijer Eric wrote: Here's what's in /usr/lost+found/ (I promise to only print this list once.) cd /usr/lost+found/ ls -l total 176014862730 b-w-r-xrw- 1 1259717448199, 228 Jun 6 2010 #85683 sr-x---r-T 1 382255288418446744073400402084 Jan 4 1966 #85688 cws--T 1 4753112439159, 161 Mar 19 2003 #85731 [...] br-srwsr-x 1 1351051690138, 231 Aug 30 2001 #87654 b--Sr-Srwt 1 4357054985 21, 26 Feb 19 1979 #87675 drwxrwxrwx 2 root root 1024 Dec 2 07:27 #98109 Ok. I like a good challenge, but this is ridicules!! This shows your disk is severly screwed up. It is time to make a backup, and see if you can reformat the disk and reinstall. If you get errors like this again, I think you will have to regard the disk as `deceased'. No way. File systems errors caused due to a crash or loss of system power cause things just like this too. It's certainly possible that this could have been caused by a power off or routine hardware failure. But this make and model of HD have a well known firmware bug. Check out, for example: http://www.wdc.com/fitnesszone/err-rec.html http://www.wdc.com/new/overlay.html Drive Firmware Quality Upgrade Updated November 11, 1996 During extensive quality testing procedures, Western Digital has identified an error recovery issue that affects less than 0.1% of a very specific group of WD Caviar hard drives. As part of our ongoing commitment to quality, we have developed a drive firmware utility that corrects the error recovery code in the WD Caviar hard drives listed below. Even though this issue affects an extremely small number of drives, we highly recommend that you upgrade your drive firmware using this utility. AC11000 (1.0 GB) AC22100 (2.1 GB) AC32500 (2.5 GB) AC33100 (3.1 GB) 0.1% is a pretty big chunk of people when you consider how popular these drives are. miket
Re: Trouble with diaresis and ssharp key
On Wed, Dec 16, 1998 at 10:57:16PM +0100, [EMAIL PROTECTED] wrote: Stefan Gundel dixit: Hello, everybody, I am presently experiencing the following problem: On my hamm installation I get only beeps when pressing the diaresis keys on my german keyboard (i.e. adiaresis, odiaresis, udiaresis). Pressing the ssharp key gives me an output similar to ESC-. . The consoles and X terminals behave similarly. Issuing commands like loadkeys ... or defining a new xmodmap map didn't improve things. On the other hand I can display documents containing these characters and can also use the keys within editors like emacs or vi. This is my /etc/.profile file configured for the Spanish keyboard (ie. dieresis, accents, n~,...). All you should do is changing: export LC_ALL=es_ES for: export LC_ALL=es_DE Shouldn't it be de_DE? Check out /usr/share/locale/*. It looks like country code (probably ISO 3166) first, then language code (ISO 639). miket
Re: Recommendations for Office...
Not sure about spreadsheets (there are several that are fine on the computational end, but nothing has the scriptability, embedding capabilities, and feature richness of Excel). Look at www.gnome.org for gnumeric; it's still pre-1.0 and a toy compared to Excel but Gnome moves _very_ rapidly and Miguel de Icaza, its author, is the head Gnome software stud: http://www.gnome.org/gnumeric/ For a word processor, check out the Linux version of Corel's WordPerfect. It becomes available from www.download.com tomorrow (December 17th) and is free for personal use and very reasonably priced for business. Haven't read any reviews of it yet, but I have heard it's a lot smaller than StarOffice (it would be hard to get much bigger). miket On Thu, Dec 17, 1998 at 12:00:33PM +1100, Austanners Wet Blue Pty Ltd wrote: Can anyone recommend a good Spreadsheet program and Word-type program for use with X. I'm looking for something that people who are familiar with MS Excel/Word could use. Pref. not Star Office (too big) Regards, Stephen Lavelle
Re: Giving recompile another try
On Wed, Dec 16, 1998 at 10:56:42PM -0600, Cristov Russell wrote: My take on RedHat vs. Debian What I like most about the RedHat distribution is that it has a setup for SoundBlaster audio cards (I also like the printer setup which was a piece of cake with my networked printer). In fact, sound cards seem to be the only hardware that does not currently have an option to setup during the installation or utility to help you configure the device afterwards. The 2.2 series kernels, which will be available soon, include the modular sound system (which was funded by Red Hat). I'm running a patched 2.1.131 here, and I was able to load a working sound system entirely from modules. I think some of the later 2.0.x kernels (.35? .36?) include at least some support for modular sound too. miket
Re: X keyboard mapping
On Mon, Dec 14, 1998 at 02:26:55PM +0100, Armin Wegner wrote: I would like to switch between two keyboard mappings - us and german - whenever I need to. I've written two input files for xmodmap, accordingly. These are correct. I typed xmodmap file but nothing happend. I can remember that this worked some installations before. What could be the fault? Are there any other methods of keyboard layout switching under X? Try disabling the XKEYBOARD extension (/etc/X11/XF86Config). miket
Re: Presort my email messages when exmh is the MUA.
On Tue, Dec 15, 1998 at 01:19:16AM +0200, shaul wrote: I am using exmh to handle my email (MUA ?). The MTA is smail (is it also the MDA ? What are the differences between MDA and MTA ?). MUA: Mail User Agent: interactive mailer examples: mutt, pine, Outlook, etc. MTA: Mail Transfer Agent: relays mail between hosts (usually via SMTP). examples: exim, sendmail, qmail. MDA: Mail Delivery Agent: delivers mail to users at a destination host. Sometimes, the MTA is also the MDA (for example, when exim puts mail directly into your mailbox with the appendfile rule). At other times, deliver or procmail might be acting as an MDA. Interestingly, delivery at the far end by the MDA, not transport across the Internet, is responsible for almost all of the time delay in sending email. miket
Re: missing header file
On Mon, Dec 14, 1998 at 10:57:21PM -0700, Robert Kerr wrote: Hi, I'm trying to port a program from HP-UX to Linux, but I seem to be missing a header file. This seems kind of strange to me, since the header file is present in the gcc directories on the HP's, but nowhere is it to be found on my LInux box. The header file is generic.h. Where can I get this? H... can't help you with this one; not familiar with HP-UX. Second question. Can someone enlighten me as to the difference between g++ and egcs, and what is compatible with what? 'g++' could be either the GNU gcc version or egcs. I have gcc 2.7.2.3 installed as my C compiler, but egcs for C++ (this is Hamm): sarcastro:~$ g++ --version egcs-2.90.29 980515 (egcs-1.0.3 release) sarcastro:~$ gcc --version 2.7.2.3 sarcastro:~$ dpkg -l | egrep 'gcc|egcs' ii egcs-docs 2.90.29-0.6Documentation for the egcs compilers (egcc, ii g++ 2.90.29-0.6The GNU (egcs) C++ compiler. ii gcc 2.7.2.3-4.8The GNU C compiler. ii libstdc++2.82.90.29-0.6The GNU stdc++ library (egcs version) Egcs (pronounced 'eggs') is hosted by Cygnus and was started because of historical concerns about gcc's slow pace of development and small, closed group of developers. Egcs _is_ gcc; it was forked off the gcc code, but I'd imagine that for C++, at least, egcs has seen some serious changes. The consensus seems to be that for C++, egcs offers much better compatibility with the ISO standard and a greater push to incorporate modern ideas in compiler design (the egcs/gcc backend is still years out of date in some areas). The way the gcc/egcs scenario will ideally work is that egcs is the more advanced, experimental compiler, and features from it are backmerged into gcc as they become stable. It has yet to be determined if this is workable. miket
Re: [exim/fetchmail] A day in the life of an email
On Thu, Dec 10, 1998 at 06:17:20PM +1030, Mark Phillips wrote: On Tue, 8 Dec 1998, Mike Touloumtzis wrote: 3) Set qualify_domain = ist.flinders.edu.au to make your locally generated mail appear to originate from that machine. This is potentially not necessary if ist.flinders.edu.au accepts mail for *.flinders.edu.au and the MX records are set up correctly in the DNS, but I don't think it can hurt. Thanks for your help. Things are almost working. Only one problem at the moment. Setting the qualify_domain to ist.flinders.edu.au works well at making my email appear to originate from uni. The only problem is when at home I email to mark, the email ends up at my account at uni. This is not the desired behaviour. mark is my local username and the email should have been sent locally. For some reason it is not. I presume that because I addressed the email to mark not to [EMAIL PROTECTED], it decided to qualify it for me to [EMAIL PROTECTED]. Is there any way of getting it to make the from address appear to be [EMAIL PROTECTED] while at the same time, allowing mark to be completed as [EMAIL PROTECTED]? Hmmm... the first thing to do is to check to see if mail for yourhostname.flinders.edu.au ends up at ist.flinders.edu.au when sent from the outside world. If this is the case, then you can just use your hostname+domain for qualify_domain and replies should work OK (this is the way my company's setup works; mail to [EMAIL PROTECTED] will get to me). Try hostname.ist.flinders.edu.au as your local host/domain + qualify_domain too, as 'ist' is not the real hostname of your smarthost--see below). I played around with 'dig' for a sec, and it looks this depends on frodo.cc.flinders.edu.au (AKA flinders.edu.au, and the catchall mail destination for your university according to the MX records) if your machine is hostname.flinders.edu. If your machine is hostname.ist.flinders.edu, it depends on ist.flinders.edu.au, AKA adam.ist.flinders.edu.au. Frodo resolves to a different IP than adam, so these are probably distinct machines. One or both of these machines might know enough about you to get mail to your spool just based on your username. 'ist' sounds like a better bet, as it's more specific to you. frodo.cc.flinders.edu.au's SMTP server accepts mail for [EMAIL PROTECTED], but that doesn't mean anything nowadays; many servers delay bounces to make it harder for spammers to verify addresses. The only way to find out for sure is to try it. Note that the success of this may also depend on whether or not your smarthost will rewrite your email address. Sendmail has some advanced anti-spam features to verify senders, and a bogus hostname might throw it. Again, the only way to find out is to try. Barring this, your best option is probably to use the address rewriting features of exim (set your real hostname+domain as the qualify_domain, but rewrite them to the smarthost for mail that leaves the machine). This is something I have never used myself, but I took a quick look at spec.txt in the exim docs, and it appears that you can turn rewriting on and off for each director rule, and specify rewriting options for both envelope and headers. There's also the command line 'exim -brw [EMAIL PROTECTED]' that will spit out the rewritten versions for testing purposes. miket
Re: [exim/fetchmail] A day in the life of an email
Hi Mark, Some answers are given below. I hope they're correct; they're based on my experience in a similar situation (although my machine is behind a firewall). miket On Wed, Dec 09, 1998 at 11:00:45AM +1030, Mark Phillips wrote: I am looking at setting up mail properly on my machine. I have an email account at university and use ppp to gain dial-up access to uni. I wish to set things up so that I can receive and send email from my local machine (rather than rlogin to a uni machine to read email as I do at the moment.) Please excuse my ignorance, but I am a bit unsure about what I must do and about exactly what happens with email. Here is my current understanding of what the story is: 1. Email arrives at the mail server at uni with the address [EMAIL PROTECTED] and is stored in a spool directory there. 2. I contact the mail server using fetchmail, asking it very nicely to remove the email from my spool directory there, and send it to my machine at home. Fetchmail does this with the POP protocol. 3. Fetchmail feeds this email into some black box (I don't understand this bit - is it a file or is it a program that is sitting around waiting for such events???). Exim somehow notices that there is email coming into this black box and stores this in /var/spool/exim/input. Exim then looks and sees that the domain of the message is ist.flinders.edu.au. Fortunately this domain has been configured as the local domain in exim.conf so exim knows to distribute these messages to local users. It then sees the user is mark and so knows to distribute this to mark on the local machine. It moves the email from /var/spool/exim/input to /var/spool/mail/mark. Fetchmail connects to port 25 on your local machine. Depending on your machine's configuration, either exim is started by inetd in response to this, or exim is running as a daemon and was already listening on that port. 4. I then run a program such as pine which sees there is email in /var/spool/mail/mark and enables me to read it. I have received my email. 5. I then decide to send a message to [EMAIL PROTECTED] I compose an email in pine which puts the message in the black box talked about above. Exim again somehow notices it and stores it in /var/spool/exim/input. It sees that banana.com is not a local domain. It decides to forward the email to some smarthost configured in exim.conf. What mechanism/protocol does it use to do this? What should the smarthost be? Should it be my home machine (I think not??). Should it be the mail server at uni? Anyway, somehow it passes the email on to some machine that somehow knows what it is doing and delivers the email. Exim uses SMTP to connect to the smarthost. For the smarthost, you want a machine with certain characteristics (see below). 6. I then send an email to [EMAIL PROTECTED] who might be a friend at uni. Again exim gets this email, but sees that the domain is ist.flinders.edu.au. I have configured this as a local domain so exim trys to deliver this to a local user. But fred is not a local user on my home machine. He is a user at uni. I want this email forwarded on to the machine at uni. How is this done? You probably don't want to your machine to have the same hostname as your university smarthost. Mail sent via SMTP is sent using an envelope, a list of addresses that is not contained within the message itself (this is what allows Bcc to work). So even though the mail that fetchmail fetches is addressed to [EMAIL PROTECTED] within the message, fetchmail uses something like mark or [EMAIL PROTECTED] as the envelope recipient when it talks to your SMTP server via TCP port 25. You don't really want fetchmail getting too smart about parsing recipients from your message, or you might end up looping the message back to all of its original recipients. Here's the way you probably want your mail to work: 1) Configure exim to treat yourhostname.flinders.edu.au as a local domain. Add whatever other local domains are needed to get fetchmail to work; this may just involve setting local_domains_include_host = true or I believe it will work to set sender_unqualified_hosts = yourhostname.flinders.edu.au 2) Remove ist.flinders.edu.au from the local_domains. This is a distinct host from yours; having your machine accept mail for it complicates the situation and shouldn't be necessary. Presumably the only thing that will ever cause mail to be routed to your machine from the outside world is fetchmail anyway. 3) Set qualify_domain = ist.flinders.edu.au to make your locally generated mail appear to originate from that machine. This is potentially not necessary if ist.flinders.edu.au accepts mail for *.flinders.edu.au and the MX records are set up correctly in the DNS, but I don't think it can hurt. 4) Set ist.flinders.edu.au as your smarthost. Smarthosts should be able to directly
Re: A few questions
On Tue, Dec 08, 1998 at 09:51:01PM -0400, Jeff Browning wrote: Now that I got X up and running, I got a few questions. 4. If I'm idle for about 10 minutes, my screen turns black, how do I turn that off or start a screen saver? To disable blanking for the text consoles: setterm -blank 0 To disable X server screen blanking: xset s off To get a (great!) screensaver for X, install the xscreensaver package and add the following to your ~/.xsession (for XDM) or equivalent for startx (.xinitrc? not sure what Debian uses): xset s off xscreensaver These lines should precede the one that execs your window manager. My .xsession is: === #!/bin/sh xset s off xscreensaver exec fvwm2 === miket
Re: emacs, PATH, trailing /
FWIW, (getenv PATH) doesn't yield trailing slashes for me with hamm and emacs20. So I don't think it's a simple hamm bug. Sorry I'm not more helpful. miket On Thu, Nov 19, 1998 at 02:54:49PM -0500, David S. Zelinsky wrote: Ever since upgrading from bo to hamm, my PATH environment variable as seen from inside emacs (with (getenv PATH)) has a trailing / on each entry: /usr/local/bin/:/bin/:/sbin/:/usr/bin/:/usr/sbin/:/usr/X11R6/bin/ This doesn't really hurt, but has the annoying side-effect that if I type `which foo' in an emacs shell buffer, I get a path name with a // in it: /usr/local/bin//foo Again, this doesn't really hurt anything (except for my sensibilities : ), although occasionally I want to take the resulting path, and paste it in a find-file command to open the file -- but then emacs replaces `.*//' by `/', and so looks for /foo instead of /usr/local/bin/foo. I would like to know what is causing this behavior. I grep'ed on PATH in all the .el files in /usr/share/emacs/19.34/lisp, but didn't find anything that sets PATH. I even made an attempt to find something relevant in the emacs source code, to no avail. I didn't see this behavior until I switched to hamm, even though it's the same (upstream) version of emacs. I also tried installing emacs20, and got the same results. Has anyone else seen this? Or have any idea what causes it? Or to whom to report a bug? -- David Zelinsky [EMAIL PROTECTED]
Re: bogomips
On Tue, Nov 10, 1998 at 10:24:46PM -0800, Oz Dror wrote: Hi I have a pentium-II 400Mhz and a pentium MMX 200Mhz both have 400.59 bogomips Why? MMX doubles your bogomips. MMX CPUs are evidently great at running empty loops. miket
Re: Command shells with scroll bars?
In addition, Shift+PageUp/Shift+PageDown will scroll in xterm, rxvt, and the console. miket On Fri, Nov 06, 1998 at 09:34:45AM +0100, Frank Barknecht wrote: Kevin Grant hat gesagt: // Kevin Grant wrote: Assuming I'm using the right terminology, I can't seem to find a command shell (like csh or xterm or whatever) that gives me a little scroll bar on the side so I can use the mouse to scroll back and see what happened before. Are there any like this? xterm shows a scrollbar, if you start it with the option $ xterm -sb which is documented in the man page for xterm. You might want to put this line into your ~/.Xresources to make xterms always show the scrollbar: Xterm*scrollBar: true Most other terminals like rxvt can have scrollbars, too. Read the friendly manuals.
Re: exim and bad Sender: header
On Sun, Nov 01, 1998 at 08:09:41PM -0800, Carl Johnson wrote: Lee Bradshaw [EMAIL PROTECTED] writes: On Sun, Nov 01, 1998 at 10:18:15AM -0800, Carl Johnson wrote: I ended up with the following fetchmail messages: reading message 1 of 31 (1088 header bytes) .fetchmail: SMTP error: 501 @kira.peak.org : colon expected after route fetchmail: SMTP error: 501 carlj : sender address must contain a domain fetchmail: SMTP transaction error while fetching from kira.peak.org I had a similar problem with messages bounced from non-existent accounts. The automatic reply from the server is from , I guess to prevent loops. Upgrading to fetchmail in slink (Version: 4.6.4-1) solved my problem. I just installed it, so hopefully it will fix my problem also. Thanks for the response. For reference: I was able to get a download to work by adding sender_unqualified_hosts = my.domain.com to exim.conf. miket
Re: how do I extract a 2.6 gigabyte .tar.gz file ? (the saga continues)
On Thu, Oct 29, 1998 at 12:39:59AM -0600, Mark Panzer wrote: Darxus wrote: I installed Windows (see what you made me do??). I installed WinZip. I told WinZip to open my 2.6gb home.tgz file... it said okay... it said this file contains home.tar, you want me to extract it to a temp dir open it ? So unlike our beloved Unix utilities, winzip CAN seek past 2gb. BUT it can't untar and unzip at the same time, and since I don't have over 5.2gb of fat32 storage space, I don't have enough room to extract the .tar that my .tgz contains to a temp dir. So I still can't get my files. (ARGH) I'm guessing Win95 can go out to 4GB 2^32 so maybe if you tried to extract this archive in Win it still wouldn't work Win95/NT have 64 bit system calls for seek() et al., so the limit is probably more than 4GB with FAT32. (BTW: how did this file get created? The whole thing would have to be over 4GB uncompressed (right?)) Well here's what you can assume, 1. The file you want is in this archive but you cannot seek past 2.0GB 2. It would be a very large effort to recompile libc for 2GB (and all of the associated programs). 3. You can access this file via M$ related utilities. The 2GB limit is not imposed by ext2fs; it's a limitation of the VFS (Virtual Filesystem) layer in the kernel. Linus has so far resisted the extension of file sizes to 64 bit on 32 bit architectures, because he doesn't like the code that gcc generates for 64-bit arithmetic. So I'm kind of curious as to how you could generate a 2GB file under Linux at all, under any filesystem. miket
Re: X and keyboard layouts
On Fri, Oct 16, 1998 at 11:42:47AM +0200, Frank Barknecht wrote: Bek Oberin hat gesagt: // Bek Oberin wrote: One of the last remaining annoyances: I use a modified dvorak keyboard layout (ie: not-quite-dvorak, but close) and do most of my work in text-mode consoles. But sometimes I have to boot up X to run Netscape or PilotManager. RedHad's X setup had -something- that automatically pulled my keyboard layout out of loadkeys and put it into X so I -didn't- have to figure out a xmodmap file (which I don't have made up) and all that stuff. I figure Debian can prob'ly do anything RedHat can do ... is there a way for this? I don't know if debian has this automatized like Redhat somewhere, but you can use xkeycaps to generate a xmodmap file for your layout. But maybe you can setup your keyboard with the basic X configuration tool /usr/bin/X11/XF86Setup or /usr/sbin/xbase-configure. I use a Dvorak keyboard, and I didn't have to do any keyboard reconfiguration under X other than the mandatory Delete/BackSpace business. I was pretty surprised it went so well, actually. I'm not using the XKEYBOARD extension, which doesn't have a good Dvorak map (I heard one had been added to the Open Group's X source, but it hasn't made its way to XF86 and/or Debian yet). If you haven't already disabled XKEYBOARD in your XF86Config, try that. miket
Re: Urgent Help needed please
On Mon, Oct 12, 1998 at 06:10:24PM +, Andy Spiegl wrote: Hi! trying to install new packages I just noticed that I can't write to /var/lib/dpkg anymore. The error I get is: No space left on device. I took a look at /var/log/kern.log and found this: ... kernel: EXT2-fs error (device 08:07): ext2_new_block: Free blocks count \ corrupted for block group 4 ... last message repeated 207 times ... last message repeated 133 times and so on. Help! What can I do to resolve this without rebooting the machine? Well, I guess I could reboot it, but it is very far away from me and if it gets stuck during the reboot I'd have an even bigger problem. Besides, there are some users logged in and I'd hate to kick them out. :-( Please send help soon! Thanks so much in advance, Andy. I've read the other messages in this thread, so I know you survived ok :-). But for future reference, here are some helpful things to know: 1) You usually don't have to reboot to fsck a filesystem, especially a non-root filesystem. First, kick off your users (shutdown -k is useful for this). Then umount the filesystem, fsck it, and remount it. This works great for /home, not so well for /var, since it tends to be in use all the time. If you can't umount it, take the system to single-user mode with 'telinit 1', then try the umount/fsck. 2) If you're wondering whether or not fsck will be run at boot time: most Linux/Unix installations, including Debian, test for the presence of a /forcefsck file in the rc scripts at boot time. If this file exists, all filesystems in /etc/fstab are fsck'ed. So: # touch /forcefsck # shutdown -r now Check out /etc/init.d/checkfs.sh for more details. miket
Re: Urgent Help needed please
On Tue, Oct 13, 1998 at 12:49:24PM +, Andy Spiegl wrote: If you can't umount it, take the system to single-user mode with 'telinit 1', then try the umount/fsck. That wouldn't work either in my case, because I only have remote access to this machine. Yeah, I somehow missed the 'remote access' bit in your previous mail :-( 2) If you're wondering whether or not fsck will be run at boot time: most Linux/Unix installations, including Debian, test for the presence of a /forcefsck file in the rc scripts at boot time. If this file exists, all filesystems in /etc/fstab are fsck'ed. So: # touch /forcefsck # shutdown -r now Check out /etc/init.d/checkfs.sh for more details. Oh, that's very interesting!!! Thanks a lot for that hint. You're welcome. I remembered after the fact that there's an equivalent /fastboot for booting without checks, too. miket
Re: segfault: yes. what?
On Sun, Oct 11, 1998 at 12:49:36AM -0400, Alexander Gutfraind wrote: Hello Fello Users! I have a small completely newbie question: What is segmentation fault AKA segfault? (I REALLY don't know) TIA for your help. I see you've gotten some other replies, but I don't think this was mentioned: a segmentation fault under Linux/Unix is the same as a GPF (general protection fault) under Windows. They almost always result from buggy programs. miket
Re: Setting size that windows start as
On Sun, Oct 11, 1998 at 11:47:29AM +0100, M.C. Vernon wrote: Dear debian people, When I launch emacs in X, it comes up just a little too big for my screen. Is there a command-line argument for setting the size of the window, please? Thanks, Matthew More info about the -geometry option (described in other replies): -geometry can usually be abbrebiated as -g - in addition to +x and +y to specify position relative to the upper-left corner, you can use -x and -y to specify positions relative to the right at bottom edges of the screen, respectively. Example: xclock -g -200-200 miket
Re: meskes@usa.net
On Sun, Oct 11, 1998 at 10:51:08AM -0500, D'jinnie wrote: : Dell offers such a configuration (about $1500 with a 17 monitor). :This is what my OM will probably get. This machine comes with windows :98 and microsoft home office software and a winmodem. If you only want :to install linux call them and ask for a quote with NO software, and a :REAL modem. The Dell dimension 333v (celeron) was rated a best buy by :PC magazine. You can't get a machine from Dell with no software at all. They won't sell it without at least Win98. If you don't feel like spending extra $100 or whatever 98 costs, you'd be better off assembling your own PC - just buy everything from the same place, they will usually put it together (esp. if you ask). Win98 costs about $80 a pop in bulk quantities, I think. If you don't feel like cobbling together a machine from scratch, there are now a number of VARs who sell pre-built Linux machines. VA Research is one: http://www.varesearch.com/ If you hunt around on the Web, you should be able to turn up a few more. miket
Trouble with SB ViBRA16X
Hi all, I'm having a problem getting my on-the-motherboard sound to work with 16-bit samples. 8 bit works Ok, but 16 bit is silent, which suggests a DMA conflict or problem. The card is a SoundBlaster Vibra16 PnP (PnP string is Creative ViBRA16X PnP). The preferred configuration from pnpdump is: (CONFIGURE CTL00f0/-1 (LD 0 # ANSI string --Audio-- # Multiple choice time, choose one only ! # Start dependent functions: priority preferred # IRQ 5. # High true, edge sensitive interrupt (by default) (INT 0 (IRQ 5 (MODE +E))) # First DMA channel 1. # 8 bit DMA only # Logical device is not a bus master # DMA may execute in count by byte mode # DMA may not execute in count by word mode # DMA channel speed in compatible mode (DMA 0 (CHANNEL 1)) # Next DMA channel 3. # 8 bit DMA only # Logical device is not a bus master # DMA may execute in count by byte mode # DMA may not execute in count by word mode # DMA channel speed in compatible mode (DMA 1 (CHANNEL 3)) ... IO ports that look fine ... The weird thing is that there are two 8-bit DMA channels listed here (as opposed to an 8-bit and a 16-bit channel). At first I thought this was pnpdump acting up on me, and I tried to configure the 2nd channel on DMA 5, 6, and 7 without success. Then I ran into a mail on the Web from someone who had this card and got it working (presumably on Win95) after solving a DMA conflict over channel 3 with the BIOS setup of his lp port. I had the same setting on my motherboard, but even with it disabled, I got bad or missing DMA channel errors loading sound.o with the second channel set to 3. So it appears that I've got a broken chipset that identifies as a SB16 but supports two 8 bit channels instead of the normal 8/16. Has anyone got this working right under Linux? Does the SoundBlaster driver for 2.0.x support it? I don't want to start mucking around in the sound code without asking a _lot_ of people first. Please, no get a real sound card responses. It's my work machine, so either I have sound or I don't. :-) thanks, Mike Touloumtzis
Trouble with SB ViBRA16X PnP
Hi all, I'm having a problem getting my on-the-motherboard sound to work with 16-bit samples. 8 bit works Ok, but 16 bit is silent, which suggests a DMA conflict or problem. The card is a SoundBlaster Vibra16 PnP (PnP string is Creative ViBRA16X PnP). The preferred configuration from pnpdump is: (CONFIGURE CTL00f0/-1 (LD 0 # ANSI string --Audio-- # Multiple choice time, choose one only ! # Start dependent functions: priority preferred # IRQ 5. # High true, edge sensitive interrupt (by default) (INT 0 (IRQ 5 (MODE +E))) # First DMA channel 1. # 8 bit DMA only # Logical device is not a bus master # DMA may execute in count by byte mode # DMA may not execute in count by word mode # DMA channel speed in compatible mode (DMA 0 (CHANNEL 1)) # Next DMA channel 3. # 8 bit DMA only # Logical device is not a bus master # DMA may execute in count by byte mode # DMA may not execute in count by word mode # DMA channel speed in compatible mode (DMA 1 (CHANNEL 3)) ... IO ports that look fine ... The weird thing is that there are two 8-bit DMA channels listed here (as opposed to an 8-bit and a 16-bit channel). At first I thought this was pnpdump acting up on me, and I tried to configure the 2nd channel on DMA 5, 6, and 7 without success. Then I ran into a mail on the Web from someone who had this card and got it working (presumably on Win95) after solving a DMA conflict over channel 3 with the BIOS setup of his lp port. I had the same setting on my motherboard, but even with it disabled, I got bad or missing DMA channel errors loading sound.o with the second channel set to 3. So it appears that I've got a broken chipset that identifies as a SB16 but supports two 8 bit channels instead of the normal 8/16. Has anyone got this working right under Linux? Does the SoundBlaster driver for 2.0.x support it? I don't want to start mucking around in the sound code without asking a _lot_ of people first. Please, no get a real sound card responses. It's my work machine, so either I have sound or I don't. :-) thanks, Mike Touloumtzis
