Re: C++ dev environment advice
Michael I'm sorry not to be more prompt in my reply to this. Some have advised using the STL -- the Standard Template (not 'type') Library. That is sound advice. Yep. [... list of documentation ...] I actually found the SGI html pages that are contained in the stl-manual debian package quite practical. It may depend on your level of understanding of C++ how much you get out of these, they are not beginner's material. Anyway, it is easy to have a look at them. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: what is ginstall?
I have had this experience with two machines running Debian. While compiling some apps, during 'make install', I get an error about not finding ginstall. I solved this by making a symlink: ln -s install ginstall Has anyone else had this problem? I take it this is not a debianized source? Often if GNU utilities are installed on non-GNU unices, they get a `g' prepended to their name, so on an SGI machine, you may find gmake (GNU make) together with make (IRIX make), and gtar (GNU tar) together with tar (IRIX tar). I suppose the app you were trying to compile has been developed on a platform that both had a native install and a GNU install. I personally would edit the makefile instead of adding links on my system, but that probably is a matter of taste. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: How to change size of page in postscript doc? (non-trivial question)
Hi All, I've got a PostScript doc, which is incorrectly formated - the usable part uses a small rectangle on each page (on each page the same, but not centered on the page), and the margines are sometimes occupied by random writings, and sometimes not. I have no sources, so I can't generate it correctly. The psresize fails to handle it correctly - when enlarging, it moves part of the usable area outside the page and leaves some rubbish on it :-(. I need an utility, where I could specify not only the change of size, but also the X and Y offset, by which the page should be shifted in the output. Does anybody know such a beast? pstops can do things like that. It requires very thorough study of the man page, and if the PS document is really shaky, your only option left is to learn postscript (it is a computer language, not just a format) and start debugging. If you can do that, you will be highly esteemed in hacker circles (but probably have lost joy in life ;). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: reading rtf files
Is there any utility/program in debian that assists in _reading_ rtf files? I only found a few that can generate them. Oh really did you? Could you please send me the names of those few... There are latex2rtf (which converts latex), enscript (which can convert plain text files) and troffcvt (for troff source). These have their own debian packages. Outside debian, I found the Ted editor, which reads and writes rtf files (http://www.nllgg.nl/Ted/). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: HP880c printing from (i.e.) GIMP
Hello, I just installed a new HP Deskjet 880c. I installed 'apsfilter' and chose the 'cdjcolor' driver. I can print from Lyx fine, but I can't seem to get nice output from the GIMP. How exactly does this work? Saving the same image and printing it in Windows (Photoshop 5) does work. There is a gs driver called `uniprint' that gives better results with a number of printers. To use it, there is a number of parameter files ending in `.upp'. Read about it in /usr/doc/gs/devices.txt.gz. I think you need `gs @cdj550.upp...' (this is what I use for a HP 870CXI). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
reading rtf files
Hi. Is there any utility/program in debian that assists in _reading_ rtf files? I only found a few that can generate them. TIA, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: less and color
Then ls is sensing its output device and works accordingly? How does it do that? With the isatty(3) function. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: New HP Printers...?
On Tue, Jan 18, 2000 at 04:48:05PM +0100, E.L. Meijer Eric wrote: Hello, I am looking at buying a new printer, and had my eyes on a HP DeskJet 710C, which I would definitely have bought if I only ran Windows. I wrote: The 710C is a Windows printer. Someone hacked together a driver that works very well for black and white for the 710, 720, 820 and 1000 Hi, I have an HP820Cse on one computer, and it's a Windows only printer as described above. I use a pbm2ppa driver, but if you print 2 pages from netscape it takes 90% of the cpu for about 2 minutes. In other words it's a real performace pig (it's a pig in windows too). In all fairness it does a really nice job, but I wouldn't buy one today. This surprises me somewhat. I found that black and white printing from linux with pbm2ppa on a 710 is really fast. I have a PII 266MHz/64MB system, and mostly print TeX output from linux. Do you have a much slower processor, or is it netscape performance that slows it down? The color version I recently tried seemed to be working, although I have to look into gamma settings and getting it to eject pages properly. This is really slow, as it is in Windows. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: hwclock --adjust in slink
I did not fully understand you. Does or doesn't the BIOS get the right time after the system is shutdown? The following happens: I boot, find the time is lagging behind, and then do a hwclock --set --date ... This sets the BIOS clock (not the system time), as I can verify with hwclock --show From the hwclock manual page I gather it is not adviced to run hwclock --hctosys on a running system, although I did it once and it worked (X11 went black for a few seconds, but it returned). I know that the hwclock.sh should adjust the clock and copy the BIOS clock to system time at boot time (the S..hwclock.sh script) so I reboot. If I reboot to windows98 first, I found that the time was NOT correct, and the BIOS clock was lagging just as much as it had been _before_ I set it with hwclock --set... My conclusion is that there must have been some process that did a `hwclock --systohc' during shutdown, but I cannot find any that does this in the /etc/rc?.d directories. In any case, there are hw K scripts: [18:17:27 /tmp]$ ls /etc/rc?.d/*hw* /etc/rc0.d/K25hwclock.sh /etc/rc6.d/K25hwclock.sh /etc/rcS.d/S50hwclock.sh [18:22:27 /tmp]$ Note that your system somehow got S instead of K for rc0 and rc6. I am running unstable. On two different systems running slink I only have the S..hwclock.sh scripts, so I guess this has changed in unstable. Having the K.. scripts run at shutdown would give the symptoms I described, but my system doesn't have them, so I don't understand what is happening. I have version 2.9g-6 of util-linux. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: New HP Printers...?
Hello, I am looking at buying a new printer, and had my eyes on a HP DeskJet 710C, which I would definitely have bought if I only ran Windows. The 710C is a Windows printer. Someone hacked together a driver that works very well for black and white for the 710, 720, 820 and 1000 (which use very similar protocols), see http://www.httptech.com/ppa/. Apparently there is an experimental color driver that is available if you subscribe to the developers list. I got this printer partly beyond my control. If I had to choose a printer, I would make sure there were good gs drivers for it available. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: hwclock --adjust in slink
[...] In any case, there are hw K scripts: [18:17:27 /tmp]$ ls /etc/rc?.d/*hw* /etc/rc0.d/K25hwclock.sh /etc/rc6.d/K25hwclock.sh /etc/rcS.d/S50hwclock.sh [18:22:27 /tmp]$ Note that your system somehow got S instead of K for rc0 and rc6. I am running unstable. On two different systems running slink I only have the S..hwclock.sh scripts, so I guess this has changed in unstable. Having the K.. scripts run at shutdown would give the symptoms I described, but my system doesn't have them, so I don't understand what is happening. I have version 2.9g-6 of util-linux. David Wright: They run all right. hwclock just precedes the random seed. So knowing I *had* to find something that runs it, I found: /usr/doc/sysvinit/README.runlevels.gz 5. Halt/reboot penultimate paragraph Then the /etc/rc6.d/SXX scripts will be executed alphabetically with stop as the first argument as well. The reason is that there is nothing to start anymore at this point - all scripts that are run are ment to bring the system down. Working out whether and when to run things like hwclock --systohc $GMT at reboot is something I've always meant to look at on a rainy afternoon. The trouble is, it's never rained enough! There I am, the rain in your life :) Great you found it. And it's a good thing to know they decided to call them K..hwclock.sh in potato anyway, for simple people like me that had the impression they understood how sysvinit works. Anyway, I think I am going to get rid of the links in rc0.d and rc6.d. They seem inappropriate for a system that is not up 24 hours a day. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: New HP Printers...?
Hello, I am looking at buying a new printer, and had my eyes on a HP DeskJet 710C, which I would definitely have bought if I only ran Windows. I wrote: The 710C is a Windows printer. Someone hacked together a driver that works very well for black and white for the 710, 720, 820 and 1000 (which use very similar protocols), see http://www.httptech.com/ppa/. Apparently there is an experimental color driver that is available if you subscribe to the developers list. I got this printer partly beyond my control. If I had to choose a printer, I would make sure there were good gs drivers for it available. Just in case anyone is interested. I looked again at it, and apparently the pnm2ppa driver that should supply color support for these printers is well on it's way to a productive state. I am going to test it soon. See http://sourceforge.net/project/?group_id=1322. Still I wouldn't really advocate buying this printer for linux as long as HP is effectively keeping the protocol it uses a secret. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
hwclock --adjust in slink
Hi all, Since the BIOS clock in my PC at home is lagging behind more than 10 minutes per week, I looked into the correction mechanism of hwclock. I am able to set the BIOS clock with hwclock --set --date, and get the expected result if I try hwclock --show. The hwclock is called in a script from runlevels S, 1, and 6: $ ls /etc/rc?.d/*hw* /etc/rc0.d/S25hwclock.sh /etc/rc6.d/S25hwclock.sh /etc/rcS.d/S50hwclock.sh All these links call the script with the `start' argument, and as far as I can tell, this should adjust the clock and the system time using /etc/adjtime. Note that if hwclock.sh were called with stop|restart|reload, the BIOS clock would be set to the system time, but no K..hwclock script exists. If I reboot after a hwclock --set ..., somehow the BIOS clock gets reset to the system time, so that when the system comes up again, the BIOS clock trails as much as it did before I --set it to the right time. This is a on a slink system. Has anyone got an idea where the BIOS clock may be reset to its previous value? TIA, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Viewing powerpoint files
Is there a utility to view MS powerpoint files? Do I have to install that monster of StarOffice for this perpose? ;^) It depends on what you call `view'. Our secretary sometimes insists on sending lecture programs in powerpoint files per email (yuck, bleah), and all that matters is the text. In that case a simple strings file.ppt | less will show you the text most of the time, after you scroll through a few pages of junk. The text is usually somewhere at the bottom. Funny thing is that you can often see older versions of the text that are invisible in powerpoint. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: libc.so.4
I recently came across a program (for which I have not found the source code, though it may be available) which seems to depend on libc4 (it gives the error message 'file libc.so.4 not found' or similar). There is no Debian package for slink containing this library, as there is for libc5. Does this mean I cannot use this program? Where would I be able to find libc.so.4? Would it hurt just putting it in /lib/? It is no longer supported in slink. I still have it on my system, I think there used to be a package in hamm. It will not hurt if you put libc.so.4 in /lib, but if you cannot find a debian package for it, it is a better idea to put it in /usr/local/lib. I think you will need to have support for the `aout' executable format compiled into the kernel as well. I would suggest the following: 1) Try to find a recent version of the program (what is it, by the way?). 2) Try to locate a hamm libc4 package. 3) Try to locate a libc.so.4 library elsewhere. Options 2) and 3) require aout support in the kernel, and option 3) requires that you manually run `ldconfig' after putting the library in /usr/local/lib. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: What is this????
I found this file somewhere: c---r- 1 8224 10280 49, 117 Dec 1 2031 fonts Can anybody tell me what that c is all about??? The c means that it is a character device, e.g. like /dev/tty or /dev/psmouse. However, since the group and user owners of the file seem unknown to your system (the numbers 8224 and 10280), it looks more like a file with screwed up attributes. You could try to run an fsck on the disk to see if it gets corrected, or try to remove it (maybe after a chattr command to change the attributes). This could be a symptom of a dying disk, but then again, maybe it isn't. When this type of problem occurs people also often start talking about scary things like file system debuggers. Could someone more knowledgeable jump in on this? This 'thing' is the reason for some trouble with apt / dpkg (something to do with xlib6g-dev) perhaps this c has something to do with that. Maybe it used to be a directory before it got screwed up? HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: CTRL-C Doesn't work??!! ....
[ about Ctrl-C having no effect ] So... in conclusion: Ctrl-C is working fine - the signal is being received. However, the shell isn't killing the running process Any ideas would be greatly appreciated. Doesn't Ctrl-C work for any program at all? It does not work for a lot of interactive programs when they are expecting input. In those cases you can usually press Ctrl-Z to stop the program, read the number in brackets that bash prints, and use that in a kill command, like this: $ xclock [ press Ctrl-Z ] [1]+ Stopped xclock $ kill %1 If you cannot kill a non-interactive program like xclock with Ctrl-C, I don't know what is going on. Note however that specific programs can choose to ignore Ctrl-C. I do not believe programs can choose to ignore a stop signal (Ctrl-Z). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Interface to NT, mount? Samba? Rumba?
- To allow Windows to natively see Linux partitions on your local machine, use the ext2fs utility for Windows 95/98, or a similar utility whose name escapes me for Windows NT. Just a remark: the utility for windows 95/98 is called fsdext2 (http://www.yipton.demon.co.uk/). It is not actively being maintained, and the versions I tried long ago are still on the web page. It provides read only access, and is `moderately stable' on some PC's, and hopeless on others. Not something I'd recommend for even slightly serious work. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: df and du disagree
David Wright said: Bear in mind that du can also seriously underreport usage when run as a user because of permissions (whereas df is always right). Besides files hiding `under' mount points, there is another type of file that is invisble to du. If a file which is opened by a process is removed, it is not really removed from the disk. As long as the process keeps the file open, it is a valid file (to that process). As soon as the file is closed, the disk space is freed again. Some programs use this fact to make sure that temporary files are cleaned up even if the program crashes: they open a scratch file, and immediately remove the link from the disk. Then the process has an open file that not even root can read (unless with a disk editor), which disappears from the disk as soon as it is closed. Sometimes this technique causes hard-to-understand filled up /tmp partitions if daemons have a leak towards a tmp file. It then helps to stop/start long running daemons to identify the problem. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: df and du disagree
Hi all, df shows my main partiton to have 1.1 Gigs of data and du -x shows it to have about 650 MB. I think du is correct. df reads FilesystemSize Used Avail Use% Mounted on /dev/hda1 1.9G 1.1G 750M 60% / /dev/hda3 15M 1.2M 13M 8% /boot /dev/hda5 1.7G 1.2G 471M 72% /home /dev/hda10387M 281M 86M 77% /var /dev/hda6 3.2G 2.6G 479M 85% /samba If anyone knows what causes this and whether or not I need to worry about it I'd be very grateful. Could it be that your main partition has data in directories that have other file systems mounted on them? E.g., is there data in /home on /dev/ha1 that becomes invisible to du because /dev/hda5 is mounted on it? HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Fortran, Lapack
On Tue, Dec 07, 1999 at 11:42:55AM +0100, Igor Mozetic wrote: How does one compile a Fortran program with Lapack library ??? I tried: # g77 File.for missing library # g77 -lliblapack File.for wrong switch # g77 -L/usr/lib -lliblapack File.for wrong switch plus superfluous option With the -l option, you should leave out the `lib' part of the name, and the extension: $ g77 File.for -llapack maybe you also need the math library (don't know if g77 does that automatically), and if you want your code to be optimized, you'd add an optimization switch: $ g77 -O2 File.for -llapack -lm Looking at your prompt `#', it seems as though you are working as root. This is not a good idea. Normally you only become root if you have to for some reason, and do the rest as an `ordinary' user. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: deleting files
On Tue, Nov 23, 1999 at 05:00:01PM +1300, [EMAIL PROTECTED] wrote:
Hi
does anyone happen to know of a quicker way of deleting 4 files out of
a directory other than the command find . -exec rm {} \;
will rm -r directory be as quick?
It will actually be much quicker (you may need to add a -f option as
others pointed out). The find command will start the rm command for
each file separately, so you will run rm 4 times instead of one
time. If you need `find' to select files on specific criteria, like
name, user, or date, you should use it together with xargs:
find . -name \*.tmp -print | xargs rm -f {} \;
This will start rm the minimum amount of times with maximum length
command lines.
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: the gimp?
On Mon, Nov 15, 1999 at 09:58:36PM -0500, Rob Mahurin wrote: On Mon, Nov 15, 1999 at 01:36:26PM -0500, Ben Collins wrote: On Mon, Nov 15, 1999 at 01:11:34PM -0500, Ian Stirling wrote: E.L. Meijer (Eric) wrote: I think it is a bad idea to call it `debian gimp'. If you do that you suggest that debian has a heavily modified, enhanced version of gimp. If I were one of the gimp developers I wouldn't like it if someone (say Corel) would take it, remove some plugins, and then tack their name on it (`Corel gimp'??). Eric They are entitled to if it's gpl software but it would be rude and inconsiderate. What about 'debian-gimp-lite'? Why does Debian have to be in there? Why not just `gimp-lite'? Why not just preach? 'gimp-libre'? Actually this is probably the best idea. Every package that has a `non-free' variant and a free one could have the free package named ..-free. This guarantees the user on the one hand that the contents of the package is free, and also provides a hint that there are non-free additions, for those that may need them. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: the gimp?
On Mon, Nov 15, 1999 at 02:08:57AM -0500, Ian Stirling wrote: Brad wrote: You do realize that if gif support were in the main gimp package, then the entire thing would have to go into nonfree, right? Since The Gimp is GPLed, Debian has every right to remove gif support and redistribute those sources to make the whole thing libre. If you really don't like that, you're free to go download the sources and compile them for yourself. I realize that Brad. Debian should rename 'the gimp' to 'debian gimp' when they take the legitimate liberty of modifying gpl software. And I did that: I fetched the pristine gimp from the gimp site. I think it is a bad idea to call it `debian gimp'. If you do that you suggest that debian has a heavily modified, enhanced version of gimp. If I were one of the gimp developers I wouldn't like it if someone (say Corel) would take it, remove some plugins, and then tack their name on it (`Corel gimp'??). Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Where's man?
On Mon, Nov 08, 1999 at 03:00:33PM +0700, Oki DZ wrote: Hi, I've been looking around for man package in www.debian.org to no avail. Would anybody please tell me the package where man resides? In my version of slink, there is a man package in the doc section of debian main. However, on www.nl.debian.org, it seems to be gone. There is a man-db package (http://www.nl.debian.org/Packages/stable/doc/man-db.html) that seems to provide the man command according to the description. Does anyone know what happened to the vanilla man package? Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: how to compile packages optimized for Pentium or Pentium-II?
On Fri, Nov 05, 1999 at 11:08:34AM -0500, Ben Collins wrote: On Fri, Nov 05, 1999 at 09:57:07AM -0600, Brian Boonstra wrote: Ingo wrote: On Thu, Nov 04, 1999 at 05:08:04PM +0100, Robert Varga wrote: How can I recompile the packages so that they be optimized for running on Pentium or Pentium-II or else? Does that mean that gcc normally is NOT Pentium optimized? No. GCC can optimize for pentiums, by default it compiles for i386 though. This is needed so that we don't produce code which wont run on some systems that we want to support. All that is need is the proper CFLAGS set (which I assume the scripts that replace the normal gcc and g++ merely add these manually, and without the need for modifying the build). Don't expect too much of these pentium-specific options. The biggest speedup I have seen for a fractal generator on a PentiumII, was the one I got with -Os. This is not pentium-specific at all. It reduces the code size, so that it fits better in the processor cache. It was only 12% faster than the regular -O2 optimization, which everybody uses. Adding -ffast-math made the program slower (sic), and -march=pentiumpro did not do anything noticable at all. I personally don't think you will note a big difference if you recompile all of debian with pentium-specific optimization options. This may change once the intel optimizations are built into gcc though. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Why use Enlightenment?
On Thu, Nov 04, 1999 at 11:33:29AM +0900, Miles Bader wrote: [...] Enlightenment seems to be as fast as any other window manager on my P133/80Meg machine, and E is a quite a bit more attractive than any other WM I've seen (I should perhaps phrase this as `most of the other WMs I've seen are really ugly'). [WindowMaker looks alright, but I find its icon/icon-dock behavior just too wacky to be usable...] Does the current incarnation of E already have a desktop pager with the same functionality as fvwm2? Some time ago it didn't, and for me it is one of the features I like most (and use heavily) about fvwm2. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Does anyone use ZIP disks to backup/restore their system? How?
On Tue, Nov 02, 1999 at 06:29:45PM -0800, John Miskinis wrote: Hi, [...] This leads me to ask if most people just backup their important files on linux, and if they lose their system, they reinstall from scratch, then restore just their important (user modified) files? Not even that. I only safeguard my own products (TeX files, fractals, programs). The rest I have on CD, and the second time I install something I usually configure things faster. Sometimes it is nice to try some new settings. This is how I always worked on Windows 95. If I had a builtin CDROM, and linux was easier to install I might opt for this, but on my Thinkpad 560, it took me 4-5 hours to get everything back. It probably would take me the same amount of time. But then again, it doesn't happen a lot. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Does anyone use ZIP disks to backup/restore their system? How?
On Tue, Nov 02, 1999 at 11:56:05AM +0100, Wouter Hanegraaff wrote: On Tue, Nov 02, 1999 at 02:11:09AM -0800, John Miskinis wrote: Hi, I'm hoping to find a set of tools that will allow a linux system to be backed up, and restored, using ZIP disks. I would really appreciate any advise on this subject, from anyone who has successfully restored a system from a multi-volume set. I've heard taper can work with zipdisks. However, if you have enough disk space you can just use tar and split to make a splitted (and gzip'd) tarfile in /tmp, and then just copy the splits to your zipdisk. [... description ...] I wrote a little program that can do this without intermediate storage: MSPLIT(1) Dividing output over multiple volumes.MSPLIT(1) NAME msplit - a utility to divide output over multiple flop pies, zip disks, etc SYNOPSIS command | msplit mount dir output DESCRIPTION This manual page describes version 0.2 of msplit. Msplit reads data from the standard input (usually a pipe), and writes it to files named output.num, where num is an increasing number, in a directory on which a removable medium can be mounted. Typical uses will be to distribute data over multiple floppies or zip disks. You should not mount anything before running the command. Msplit will ask the user to insert the removable volume, and try to mount it with the command `mount mount dir'. Then it will open a file named output.0, and write until there is no more input or the volume is full. If the lat ter case applies, msplit will umount the directory, ask for another volume (usually a disk), mount it, and start writing output.1. This process continues until there is no more input. Note that msplit does not erase any data found on the disks, it only adds to them. This makes it more flexible than using regular split(1). Another advan tage above split is that no intermediary files are cre ated. The current version also lists the contents of the mounted volume, and offers the possibility to erase everything. Restoring without intermediate files could also be done with a procedure like this (and this applies also to archives made with regular split): Open an xterm, and do $ mkfifo tarfile $ tar xvf tarfile Open another xterm, and supposing you have used msplit and generated files named backup.0 .. backup.2 on different zip disks, do something like (this assumes you use bash) $ exec 3tarfile insert first zip disk $ mount /zip; cat /zip/backup.0 13; umount /zip insert second zip disk $ mount /zip; cat /zip/backup.1 13; umount /zip insert third zip disk $ mount /zip; cat /zip/backup.2 13; umount /zip $ exec 31 Be sure to do this from the right directory. In between the cat commands you can give any commands you like, but make sure to cat the backup files to file descriptor 3 in the right order! If I get around it, one day I will write `mcat' which automates this procedure and complements msplit. Personally I use msplit mostly for floppies, so it is in general not impossible to store the intermediate files. If anyone is interested in `msplit' I can email it to them. If anyone feels like putting it in a debian package I'll tack on a GPL licence and you can go ahead. Of course there are no warranties ... -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: C programing
On Tue, Oct 26, 1999 at 05:11:45PM +, John Carline wrote: Ingles, Raymond wrote: On Mon, Oct 25, 1999 at 10:07:28PM +, John Carline wrote: [...] After spending the last two days trying to convert a C program I wrote some 6 years ago in microsoft C into linux. I just have to echo this question. This sort of question pops up on the C newsgroups and mailing lists a lot. There *is* no ANSI C way to, for example, clear the screen. The DOS functions from conio.h are completely separate from the C standard. (Indeed, they are heavily tied to real-mode 8086 and CGA/VGA graphics and PC hardware, and often can't even be *emulated* well on other hardware.) Yup! you've just described my plight. Naturally my code is *heavy* in calls to conio.h and graph.h. For simple plotting to a window, be sure to look at libplot from the plotutils package. It is rather basic, but easy to use if all you need is a simple plot. As a bonus, it has the same interface for plotting to a window as for plotting to postscript and various other file formats. Plotutils has info documentation. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: C programing
On Mon, Oct 25, 1999 at 10:07:28PM +, John Carline wrote: cut snip [cut again ...] After spending the last two days trying to convert a C program I wrote some 6 years ago in microsoft C into linux. I just have to echo this question. Is there no linux specific/best book that covers gcc and g++. One that includes all the standard library calls . I currently have four books on C (not the Kerninghan book though. I'll have to go look at it) and they're basically worthless. I'm not sure if it's that they're simply too old or too 'microsoft', but I'd love to find a book on gcc that would be a simple but complete reference for the occasional C programer. [ standard comment: please limit your line length to = 72 characters ] For a good general overview of the standard C library install the glibcdoc debian package, and type `info libc' (or use your favorite info reader). It is quite readable and has a lot of info on programming with ANSI standard C library functions and on typical unix programming stuff like pipes, regular expressions, sockets, process control. If you need to control character input from the keyboard and position the cursor in a terminal window, then install the ncurses development package and read the man page (man ncurses) which is rather elaborate. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: How to shut up dselect?
On Thu, Oct 21, 1999 at 11:28:15AM +0200, Daniel Haude wrote: Hi, I have a question related to installing non-debian programs: When I first installed Debian 2.1, I noticed that it came with teTeX 0.9. I un-installed that and installed teTeX-1.0 from the CTAN archive. Of course, the debian package manager doesn't know about this, so whenever I use dselect, it comes up with this old some package needs teTeX line. I once even didn't notice and let dselect have its way, so all of a sudden it started installing teTeX 0.9. I hit Ctrl-C to stop that (I know, bad idea), and now teTeX 0.9 is so fucked up that dselect won't even un-install whatever fragments it managed to put on the disk. It tells me to install it first and then un-install it. In this case, this is not a big problem because I have teTeX entirely under its own tree in /usr/local/teTeX, so messing around with another version will not do any damage. But if I had installed teTeX-1.0 in the same place as 0.9, it would now certainly be broken. You should _always_ install non-debian packages under /usr/local. That is the only way to guarantee that debian will not fudge with it. Practically all compilable software you find in the wild will by default install itself in /usr/local anyway. Is there a way to tell dselect: I installed sucha-and-such myself, it's there, so stop bitching (and remember next time)? I know that this is somewhat against the whole idea of packet management. Is making a .tar.gz into a .deb package the only clean way? What if I don't want to make the .tar.gzipped source tree the packege, but the make installed result (with all its files scattered in various places of the system)? The equivs package was designed to deal with this situation. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: reg exp question
On Wed, Oct 20, 1999 at 10:56:12AM +0100, Oliver Elphick wrote:
Shao Zhang wrote:
Hi,
I have very limited reg exp knowledge, but I need this badly...
Could someone tell me how to write a sed script to replace the
following string?
From string \textsc{Foo} to Foo
Thanks for your help.
sed -e 's/\\textsc{\(.*\)}/\1/'
This goes wrong if you have two of these expressions on one line:
\textsc{Foo} \textsc{Bar}
gives
Foo} \textsc{Bar
If `Foo' cannot contain any `}' characters, then you can solve this with
sed -e 's/\\textsc{\([^}]*\)}/\1/g'
If `Foo' does contain `}' (like \textsc{\Foo{\Bar}}) then I don't know
how to handle that.
If this really is TeX, you could probably add
\def\textsc#1{#1}
somewhere at the top of your document to get the desired result.
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: printing specific pages
On Wed, Oct 20, 1999 at 01:11:38PM +0200, Jean-Yves BARBIER wrote: Hi all, I did not found anything about printing only certain pages. Is it possible? (I'd like to print odd pages, then even on the forms verso) I mean from the command line, not from a word processor. If the postscript file is well-formed (which many are not), you could use psselect from the psutils package. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: hda: irq timeout
On Mon, Oct 18, 1999 at 04:58:56PM +0200, Peter Weiss wrote:
Hello,
running kernel 2.2.12 on a new machine I found the following kernel
messages in the syslog file with decreasing time intervalls:
Oct 18 16:28:38 Winona kernel: ide0: reset: success
Oct 18 16:30:45 Winona kernel: hda: irq timeout: status=0xd0 { Busy }
Oct 18 16:30:47 Winona kernel: ide0: reset: success
Oct 18 16:33:15 Winona kernel: hda: irq timeout: status=0xd0 { Busy }
Ouch!
[...]
Does anybody have ideas/ suggestions?
Make backups quickly! I have seen this kind of messages three times
now, and each time they came from a dying hard disk. Especially the
fact that it occurs at decreasing intervals is alarming. Try to backup
valuable data first, and then do a fsck on the disk. If you find a lot
of errors, it is almost certain the disk is failing.
The problem could also be in the hard disk controller or the
motherboard. Maybe reseating the connectors helps.
Good luck,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Netscape and its cache
On Mon, Oct 11, 1999 at 03:22:12AM -0400, Arcady Genkin wrote: Hi all: I specified in Netscape's preferences disk cache of 1000kBytes. However, whenever I go and du ~/.netscape/cache, I invariably get some huge number: 3,5Mcache And this is because I deleted its contents manually several days ago. Now, even funnier. When I do Clear Disk Cache from the same Preferences dialog, it only clears a small fraction of the space used. Here's what I saw after clearing: 2,8Mcache So it only cleared 700K. This pisses me off. Any ideas, anyone? I noticed the same thing. It is broken as far as I can see. I added the following line to my crontab: 50 ** 0/bin/rm -rf /home/tgakem/.netscape/cache This completely removes the cache once every week. After I start netscape when the cache dir is gone, it pops up some error window, I click OK or something like that, and continue with a neatly cleaned disk. It's not ideal, but workable for me. I don't belong to the crowd that gets annoyed if they have to restart netscape every week. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: file with prz extension
On Tue, Oct 05, 1999 at 11:22:22AM +0200, Istvan Benak wrote: Hi all! How can I view with my Linux box the somfile.prz file? Which program should I use? Or if someone have a program which can view this file and can convert it to an human format (i.e. pdf, or ppt, or anything else) please send me a mail Does the `file' command recognize it (type `file somfile.prz')? Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Language of www.debian.org.....
On Tue, Sep 28, 1999 at 06:11:15AM -0700, Martin Waller wrote: er - who's been hacking into the Debian www site? The main page appears to be in elvish or something... ??? The main page of www.debian.org looks OK to me. The Dutch mirror www.nl.debian.nl however, seems to be in Turkish by default. If you go to the bottom and click on `Yngilizce', you will get an english version. Now I don't know how the default changed... HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
[OT] cdrom speed adjustment
Hi all, I know someone who has a windows program that limits the speed of his cdrom drive. Is there a way to do this in linux? Although it is a nice idea that my ide/atapi drive can do 36-speed, it sounds like a hoovercraft taking off when it runs that fast, and for most applications 8-speed or something like that would be plenty. It also seems to me that it actually takes a lot of time to get upto speed, so that it sometimes _feels_ slower than the trusty old double-speed drive I used to have in my old 486. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: [OT] How to find the exact time, when the serial data arrived?
On Fri, Sep 17, 1999 at 10:14:28AM +0200, Wojciech Zabolotny wrote:
Hi All!
I'm writing an application requiring the exact knowledge of arrival time
of serial data (resolution below 0.1 sec is desired).
Is there any way to arrange it in the standard Linux kernel, or should I
modify the serial driver? (Or even use the RT Linux?)
You can use gettimeofday, which returns the time in a struct:
struct timeval {
long tv_sec;/* seconds */
long tv_usec; /* microseconds */
};
(see the man page)
I think the real resolution of this is 0.01 seconds. If that is not
enough, there is high resolution timer device you can compile into the
kernel. If you look for it in the kernel configuration (make xconfig or
equivalent) you should be able to find it.
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Which package contains:
On Fri, Sep 10, 1999 at 07:49:32AM +0200, Urban Gabor wrote: Hi, Which package contain the following programs: xload xcontrib xinfo Hmm, I don't know this one, there is one in the non-free scilab package, but maybe that is not what you mean. What does it do? I do know xdpyinfo, which is in xbase-clients. xman xcontrib or should I compile them from source? :-) For questions like this, you ftp to your local debian mirror, go to the stable (or unstable, if you like living dangerously) directly, and pick up the Contents-i386.gz (or Contents-alpha.gz or whatever) file. Then you do zgrep mycommand Contents-i386.gz or zgrep '\mycommand\' Contents-i386.gz and you will know in what package the command is. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Why use Debian? Why not Red Hat?
On Wed, Sep 08, 1999 at 07:54:49PM +0200, [EMAIL PROTECTED] wrote: I don't wanna start a flame war, but it is reeelly all that difficult to use 7(or is it 9? -I forget) installation disks instead of two?. Well ... I remember doing an install on a system with a dodgy disk drive. Making the install floppies to find out during 3 attempts that floppy 2,3, and 6 need to be written again is somewhat annoying. I have always found the combination linux/rawrite2/floppies _much_ less reliable than one might desire. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: ostrstream question (OT)
On Thu, Aug 26, 1999 at 10:03:21PM -0600, Robert Kerr wrote: Hi all, I'm having some problems converting my app to run under linux. It runs alright under SGI, HP-UX, Solaris and WinNT, but it crashes beautifully under linux. Anyway, I have a member variable called journalString of type ostrstream. I instantiate it so: journalString = new ostrstream(); journalString-rdbuf()-setbuf(NULL, 50); but when I try to use the operator with it, it Segfaults. Example crashes: *journalString ends; *journalString s ' '; where s is a const char * *journalString name; where name is a const char * Am I using ostrstream wrong? Any ideas? It looks to me as if you are doing too complicated things. Is the rdbuf-setbuf(NULL, 50) meant to allocated memory? You don't need to do that, the class is designed to take care of that by itself. Is there a reason to have a member ostrstream pointer, instead of just an ostrstream? Typical use would be something like ostrstream* journalString = new ostrstream(); string s = something; *journalString bla ' ' s ends; journalString-c_str() returns the buffer as a C style string (char*). If you use that, you need to deallocate the memory pointed to by c_str() yourself, unless you later call journalString-freeze(0); Note that in the future the ostrstream will be replaced by ostringstream. Finally, segmentation faults can be due to memory corruption happening elsewhere in the program, and only triggered in your ostrstream lines. The fact that a program does not crash on other platforms does not mean there are no bugs. It is well possible that a memory corrupting bug only bites you in linux. HTH, Eric BTW, this is more of a question for a C++ newsgroup, like comp.lang.c++.moderated. -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: fvwm2 configuration
On Tue, Aug 24, 1999 at 10:37:35PM -0500, Brian Servis wrote: *- On 25 Aug, John Carline wrote about Re: fvwm2 configuration E.L. Meijer (Eric) wrote: [ problems to get a pager with multiple desks in fvwm2 ] Probably the simplest way would be to copy the system.fvwm2rc into your home directory as .fvwm2rc and then edit it so that it puts up exactly the FvwmPager setup you want. You then wouldn't need to put it into the post.hook. That or comment out the items in .fvwm2rc and leave your post.hook as is. The system.fvwm2rc in Debian's setup is used to read all the system and user .hook files. It is designed to be robust enough to allow the user full configuration of the wm or no configuration at all and just use the defaults. Yup, I did not want to break that system. [ ... ] I use the following at the top of my ~/.fvwm/post.hook and it works great. ButtonStyle Reset DestroyDecor default-decor DestroyModuleConfig FvwmIconBox DestroyModuleConfig FvwmButtons DestroyModuleConfig FvwmPager I then have a clean slate to define my own setup for the above modules, etc. This is not the point. The number of desks cannot be configured this way, it is set when the FvwmPager module is started: FvwmPager 0 0 This is the default, running from desk 0 to desk 0 (that makes a total of, er, 1 (one) desk). I got a private email from Michael Tempsch, in which he pointed me to the documentation in /usr/doc/fvmw2/README.sysrc.gz. There it says: Configuring your pager You can put *FvwmPagerXXX lines in the post hook to reconfigure the pager to your liking: your options will override the defaults provided. This is good enough for most people. Unfortunately, you can't easily change the number of desks covered by the pager. This is because the pager is started just after the post hook. If you want multiple desks, you could wait for the appearance of the pager, kill it, and start a new one configured to you liking. You can do this by including something like the following at the end of your init-restart hook: + I Wait FvwmPager + I KillModule FvwmPager + I Module FvwmPager 0 3 This is what I was looking for. Thanks, Michael! (I hope you don't mind me telling about private email on the list...:) Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
fvwm2 configuration
Hi all, I use fvwm2 as a window manager in slink. In /etc/X11/fvwm2/system.fvwm2rc, the FvwmPager module is started with one desk, after .fvwm2/post.hook is read. I like to start FvwmPager with three desks, and do that in my post.hook file. As a result there are always two FvwmPagers running, on top of each other. Luckily the one I added is on top most of the time, but sometimes it isn't. Is there a way to make fvwm2 stop processing the configuration files in .fvwm2/post.hook? I tried to play around with the KillModule command, but that seemed to have no effect. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: gs/gs-aladin, (e)pswrite and setlinewidth
On Mon, Aug 23, 1999 at 01:03:59PM +0200, Clemens Heuberger wrote: I have the following problem: I have a postscript-file which draws to lines, a thin one and a thick one: thue:~/test $cat simple.ps %! newpath 0 20 moveto 100 0 rlineto stroke newpath 0 50 moveto 100 0 rlineto 10 setlinewidth stroke showpage I would like to have it rewritten as a eps-File (Of course, I could do this myself in this small example, but actually, I have a much bigger file which contains output from latex+psfrag and all kind of things which I want to get rid of), so I use epswrite: thue:~/test $gs -sDEVICE=epswrite -sOutputFile=simple.eps -dBATCH -dNOPAUSE simple.ps GNU Ghostscript 5.10 (1998-12-17) Copyright (C) 1997 Aladdin Enterprises, Menlo Park, CA. All rights reserved. This software comes with NO WARRANTY: see the file COPYING for details. In the result, the thick line has become thin. (I am using gs on slink; I also tried gs-aladin from slink and potato, same effect ...) Any thoughts? Tried it. Same result. Then tried ps2epsi simple.ps simple.eps This seems to work OK, it is also in the gs package in slink. It looks like a bug in the epswrite device of gs, worth of reporting. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: debian installation
On Mon, Aug 23, 1999 at 09:26:46AM -0300, JARDINE, Jeff wrote: I think you're absolutely right. I'm still working on my first installation (2 months and counting). Linux is *not* happy with a PnP soundcard and CD-ROM. From everything I've read, it appears to be necessary to recompile the kernel when configuring PnP hardware. Unfortunately, this can not be done when you only have the base kernel installed. So, it seems, first-time installation from a PnP CD ROM is impossible. Some BIOSes have an option to configure PNP devices. On my asus P2B for example there is an option `PNP aware OS', or something to that effect. If I set that to `no', I can use my PNP soundblaster without isapnptools, because the BIOS takes care of the initialisation. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: C++ compiling problem
On Thu, Aug 19, 1999 at 11:09:35AM -0600, Robert Kerr wrote: Okay, here's the situation. I'm porting a large project to Linux. Some of the libraries we have to use were compiled with an older version of egcs (I'm pretty sure it was 1.0.3). So, they expect the libraries (specifically libstdc++) that come with that version of egcs. Okay, so I installed the older version of egcs from Debian 2.0. I installed it by hand so it wouldn't overwrite my current version of everything. I set different path names and all that. Anyway, part of our code, which egcs 1.1.1 handles perfectly, causes a compiler crash for 1.0.3. H. So, here's what I propose to do: I've downloaded a copy of libstdc++2.8.0. I am planning on compiling it using the egcs 1.1.1 compiler. Then I should have the new compiler with the old set of libraries. Does anyone see any major problems with this idea? It seems a lot of trouble. Why don't you just install libg++2.8.2-dev 2.91.60-5 (slink)? This contains the parts of libstdc++ that were gnu extensions and are dropped in favour of stl and friends in the newer versions. I think all you then need to do is after installation add a -lg++ flag when linking. It seems a good idea though to try to convert the code if you continue using it in the future, since libg++ is no longer actively maintained. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: bash and parameters
On Sat, Aug 14, 1999 at 12:50:20AM -0500, Lance Hoffmeyer wrote: When I create a shell script how do I pass parameters to it? For example, if I want to create a directory based on a name I pass to the program with a shell script called mkmine the command would look like mkmine Mydir and this would create a directory called Mydir would the script simply be mkdir %1 ? No, that is a DOS batch script. bash is really different, it uses $1 :) if I wanted it to create a directory based on a name I give it and the current month would it be mkdir %1 date %m ? No again, this is DOS-speak. A complete bash script to do what you want looks something like this: #! /bin/bash mkdir $1`date +%m` The first line ensures the script is executed by bash. The backquotes around `date +%m` return the result of the date command as a string. An alternative way to get this result is mkdir $1$(date +%m) You make the script executable with chmod +x script HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: [OT] How fast is a PIII?
I wrote: Does anyone know how fast is a PIII, say 500MHz? How does it compare to PII or Celeron? http://www.zdnet.com/pcweek/stories/news/0,4153,391810,00.html says it's only 8% faster than PII at same clock speed when running business apps on Windows. If anyone has Linux benchmarks, post 'em! The only difference between a PII and a PIII at the same clock speed that I know of are the extra `SSE' instructions that are mainly useful for 3D stuff and maybe some other floating point intensive software. As far as I know, there are currently no compilers for linux that actually use these instructions, so I would expect the difference is about 0%. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: [OT] How fast is a PIII?
The only difference between a PII and a PIII at the same clock speed that I know of are the extra `SSE' instructions that are mainly useful for 3D stuff and maybe some other floating point intensive software. IIRC, the 2nd level cache of the Pentium III is running at the same speed as the CPU, wheras the 2nd level cache of Pentium II's is running at half the CPU speed. No, the PIII has its 2nd level cache running at half processor speed. The 2nd level cache does run at full processor speed in the Xeon and PIII Xeon processors, which are ridiculously overpriced. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: bash manual/info lacks examples
Says Mike: bash shells can do: let a=$b+$c*$d let a=($b+$c)*$d but if you want to ensure compatibility with Bourne shells like ash, you should stick to: a=$(($b+$c*$d)) a=$((($b+$c)*$d)) but if you want to ensure compatibility with Bourne shells like the Bourne shell :), you should stick to a=`expr $b + $c \* $d` a=`expr \( $b + $c \) \* $d` to complicate things, the spaces in the expression after `expr' are required. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: bash manual/info lacks examples
It's been a LONG time since I've use the authentic one-and-only Bourne shell. How many Bourne shell clones do we have floating around here these days? Mike I don't know about clones, but I am typing this on an SGI box where sh _is_ the Bourne shell, and I also have access to a Solaris server where this is true. OTOH I don't think there is a real problem using ksh, ash, zsh, or bash features, unless the first line of your script reads #! /bin/sh HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Sound error: Couldn't allocate DMA buffer
Look at the May 1999 archives for this list, with the subject Can't allocate DMA buffer (as opposed to Couldn't!) for a clearer explanation than I could give. Hm, now I check the whole archive of 1999 and can't find it. Could it have been on a different list maybe? No, it is there, I checked it. Note that the archives for each month are divide into pages, which you can browse from the top of each page. The URL of the page on which the discussion is, is http://www.nl.debian.org/Lists-Archives/debian-user-9905/thrd2.html HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Any tool to write/read/convert 16bit TGA files ?
See if 'convert' in the imagemagick package will do. Note that this is non-free, so you need to check the license also. convert is part of the imagemagick package, which is in main. The license of imagemagick is very free. There are some related libraries that also have non-free versions, probably those handling GIF and friends (due to the famous compression patents). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Flaming Debian Newbies
Nathan == Nathan E Norman [EMAIL PROTECTED] wrote: Nathan On Tue, 27 Jul 1999, E.L. Meijer (Eric) wrote: Nathan : Since many people who are slightly familiar with any form of unix will Nathan : be likely to try the man command on a new debian system, it may be a Nathan : good idea to include a script called `man' in the base system that Nathan : prints a message explaining that the man pages are not yet there, where Nathan : to get them, how to install them, and where to find other sources of Nathan : information (/usr/doc, ww.debian.org, this list). Nathan AOLAgreed/AOL Nathan Why hasn't this been done? Martin wrote: File a wishlist bugreport on boot-floppies. You have to indicate your wishes, the developers can't think of everything :-) Um, they were giving me that impression though :) OK, I will file a wishlist bugreport on boot-floppies. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Flaming Debian Newbies
Lee Elliott wrote: [...] Having said that, when I followed the instructions (from Debian.org) for installing slink on my platform and finally got it to load, and logged in, I found that I didn't have 'man'. It took more RTFM'ing and figuring out the various things dselect was telling me before I tried looking for a discrete 'man' package to D/L and install. Ed Cogburn wrote: [...] The man package is far to large to fit on the base system. I had exactly the same problem when I installed Linux for the first time. 'man' was one of the few unix commands that I knew (thanks to a book I bought), but after installing it wasn't there. Ouch. Since many people who are slightly familiar with any form of unix will be likely to try the man command on a new debian system, it may be a good idea to include a script called `man' in the base system that prints a message explaining that the man pages are not yet there, where to get them, how to install them, and where to find other sources of information (/usr/doc, ww.debian.org, this list). Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Programming question: sizeof struct?
Joop Stakenborg writes:
Hi there,
The upstream maintainer of one of my packages is having problems
with his code. I thought it would be nice to use the debian mailing
lists, to see if we can an answer on this. I will forward any solution
to him.
--
The reason why I have not released LogConv 1.54 for Linux is that I am
having problems with packed structures that is causing some file formats
to not be handled properly. Even though I specify -fpack_struct the
generated code does not appear to actually do this. Structure fields
are
offset and the return from sizeof() returns a value that is not valid.
The return value from sizeof _is_ valid. It returns the amount of
memory the object occupies in memory. Maybe you don't like it, but
that is another issue.
For instance, if the structure were:
struct foo {
char text[3];
int num;
};
sizeof would return 6 and not 5.
Um, I get 8 instead of 7 (an int is usually 4 bytes in linux):
oef.c
#include stdio.h
struct A {
char a[3];
int b;
};
int main(void) {
printf(%u\n, (unsigned) sizeof(struct A));
return 0;
}
If I compile and run like this:
$ gcc -O2 -o oef oef.c
$ ./oef
8
... I do get padding, and if I compile and run like this:
$ gcc -fpack-struct -O2 -o oef oef.c
$ ./oef
7
... I don't get padding. This is using gcc 2.7.2.3. I can imagine
that there exist optimization options that affect this though,
especially the -malign-... type. Maybe you need to switch these off.
There is a second, very-unportable-too-I-guess way:
oef2.c:
#include stdio.h
struct A {
char a[3];
int b __attribute__ ((packed));
};
int main(void) {
printf(%u\n, (unsigned) sizeof(struct A));
return 0;
}
This yields:
$ gcc -O2 -o oef2 oef2.c
$ ./oef2
7
However, unless you are really tight on memory, it is a bad idea to muck
around with the default padding in structures, since it degrades
performance. If the issue is to write binary files that are portable
across different architectures, then padding in structures is only one
of the problems you will encounter. Have a look at the C-FAQ-list at
URL: http://www.eskimo.com/~scs/C-faq/top.html, item 20.5, for
references on how to tackle this problem.
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: A Pet Peeve about posting on the lists
Quoting Stephen Pitts([EMAIL PROTECTED]): SOAPBOX If you wish to start a new thread on the list, PLEASE send a message to [EMAIL PROTECTED] DO NOT just reply to any old message in your mailbox. If you do that, then the threading gets messed up, and it doesn't display right in the archives, or in threaded mail clients like mutt. [...] Might it be that these mail programs have poor docs? Might it be that the people that use them don't realize how badly these posts mess up the display of good MUA's. Might it be OK for me to recommend that they take a look at mutt. Maybe it would be a good idea to put some `posting guidelines' in the email that confirms subscription to the list. Some of my own peeves for that list would be: *) Use a descriptive subject, not `Linux problem', or `help needed', or `Unidentified Subject!', and certainly not `HELP, URGENT!!!', but more something like `Debian won't boot', or `Lost mail with elm'. *) Read the list before responding to it. If you know the answer to a question, first check if it has not been answered before. The list is busy enough as it is without twenty explanations of how to delete a file named `-r -f *'. *) Quote as much as necessary, but not more. Please put your comments _below_ relevant quoted text, not above. *) Remember people are taking the trouble to reply to your questions in their own time. If your question isn't answered the first time, try a different subject line, don't get angry. There is no plot to ignore you. *) Information for new users can be found at ... Newbie Guidelines ... /usr/share/doc ... etc ... HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: How to switch off line buffering in stdin?
Hi All! I'm writing an application, which implements some terminal functionalities. I'd like to receive every keystroke, just after the key is pressed (like with vga_getkey(), but in text mode). The standard fgetc(stdin) receives the char only after the whole line is entered. The setvbuf(stdin,NULL,_IONBF,0); doesn't help at all. How to implement it? Probably I should use ioctls to change the console's behaviour. Where should I look for the information? You will want to use the curses library: man ncurses (yes that has an `n' in front). To compile programs with ncurses, you need the ncurses development package, libncurses4-dev. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
fstab entry for floppy
What do I put in my fstab to be able to mount the floppy with just 'mount /mnt/floppy'? Currently I have a line, but that one makes the system try to mount it at boot, which fails... If you add `noauto' to the options on that line, the system will not try to mount it automatically at boot time. You also may want to add the `user' option to allow every user to mount the floppy (not just root). HTH, Eric P.S.: It is a good idea to give your emails on the list a short but descriptive subject. -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Free Financial Program (Somenthing like DAC-EASY or QuickBooks) ???
I need a free Financial Program (Somenthing like DAC-EASY or QuickBooks) I will prefer if the program is an X Window aplication, but I will be glad with a console application I don't know the apps you talk about, but you could look at GNU Cash (which I never used myself). See http://www.gnucash.org/. I don't think it is debianized (yet?). Anyone with experience with this package? HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: suid question, kind of
I forgot how to make a program start when the machine boots, but not have it start as root. I want it to start as another user. Any ideas, anyone? Look at `su username -c ...' or setuid from the super package. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: package dependency
Hi, How do I know if the package is no longer dependent by others? for example, if I have a lib-blah package installed, and I want to know if there are any other packages installed on my system that depends on this package. I don't want to try to remove it though. This information is available in /var/lib/dpkg/status. You can get it out with the little sh script I called `debquery' which is attached below. I donate this sophisticated piece of software engineering to the debian community :) You can ask for any field you like: # default is the Depends field $ debquery xlib6 xaw95 xview tk42 xquake lesstif # other fields are possible: $ debquery Suggests gnuplot octave `debquery Status installed' would give you all installed packages, but, unlike `dpkg -l' not cut off longer package names. You need /bin/sh and `ex' installed for the script to work. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA) -- Cut here and save as `debquery' - #! /bin/sh # # This software may be distributed under the GPL. # I like to call it `debquery' # # (C) Eric Meijer, 1999 # if [ $# -lt 1 ] || [ $# -gt 2 ]; then echo usage: $0 \[\field\\] \debian package\ echo field can be one of the following: echo Depends (the default), Suggests, Conflicts, or any other echo field found (or even not found :) in the status file exit fi if [ $# = 1 ]; then FIELD=Depends PACKAGE=$1 else FIELD=$1 PACKAGE=$2 fi STATUSFILE=/var/lib/dpkg/status ex $STATUSFILE _EOF_ | cut -d: -f2 g/^$FIELD:.*\$PACKAGE\/ ?Package? p _EOF_
Re: package dependency
[ my `debquery' script ] Right, this is the first update... %) [...] ex $STATUSFILE _EOF_ | cut -d: -f2 g/^$FIELD:.*\$PACKAGE\/ ?Package? p The previous line works better if it replaced by these two: g/^$FIELD:.*\$PACKAGE\[^-]/ ?Package? p g/^$FIELD:.*\$PACKAGE\$/ ?Package? p This prevents `gs' from matching `gs-aladdin'. Hope that's it in the bug department, since I'm leaving for a few days. Eric (feeling that regular expressions are always more hairy than they look). -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: User menu problem
Hi all, I have just installed StarOffice 5.1 in my Debian 2.1 system. I tried to put an entry into the menu system (I use fvwm2 as a window manager). So I created a file .menu/soffice in my home directory with the contents: ?package(local.soffice):needs=x11 section=Apps/Editors title=StarOffice \ command=~/Office51/bin/soffice and then I run update-menus command. Unfortunately there is no new entry in the menu. What did I wrong? Thank you for any suggestions. You also have to restart the window manager. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Help find man pages for Libc functions +
Hi, I once had a problem moving from bo to slink. Many packages were removed and had to be reinstalled. One of the things that got lost is the documentation on C functions. For example I used to run man strcpy and the documentation of the function strcpy will be displayed, now it does not work. Obviously this means the man pages are not installed. My question is which package do I need to install? The C-library function descriptions are in the manpages-dev package. The info pages in the glibcdoc package are also very useful to a C(++) programmer. The second problem has to do with lpd. It seems to die once in a while. If I want to reinstall it which package should I install? I have the same problem with lpd from lpr_0.33-3.deb. Installing the last version did not help me a lot. You might want to try the lprng package (maybe you even have that installed). This should have enhanced functionality. I never got around getting my print filters to work with it though (yeah, I know about magic filters, but since I wrote my own in my old slackware days, I kept it that way). For now, I just restart lpd once in a while if I need to print. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: pbm2ppa printer filter
I was just poking around to see if anyone has gotten pbm2ppa working under debian. i had it working under rh5.2 a while ago... Yes, I have it working, but not it is on my home PC, which is unconnected to the internet. The author has developed it on debian as I understood. There is a mailing list for pbm2ppa (_very_ quiet compared to this one), that recently had a short thread on pbm2ppa on debian. Tim Norman posted a magic filter for use with pbm2ppa, see the archive at http://www.listbot.com/cgi-bin/view_archive?Act=view_archivelist_id=ppa.users HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: V. basic script question
I'm on a dial on demand connection and I have this script in /etc/ppp/ip-up.d/ for having the mail fetched from my ISP everytime I make a connection, and then sent locally: - start script - #!/bin/sh /usr/bin/fetchmail -f /etc/fetchmailrc -a -u a4608456 /usr/sbin/sendmail -q - end script - now, this works alright but I'd like to know when everything is over with a message echoed on screen... how do I go about that? I don't know exactly what you mean by `echo on the screen'. The most basic thing would be to print in the terminal where the script runs: echo Ready. If you want a window to pop up, you might use xmessage from the xcontrib package: xmessage It\'s done. Yet another way would be xterm -e sh -c 'echo All is done; read a' The `read a' portion is used here to wait for someone to hit enter. If the ppp script is not running as the user that runs the X11 session (I don't know, never used ppp), then you may have to put something like xhost +localhost in your .xsession file to allow the window to pop up. HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: how do i produce a core file?
Quoting Jeremy Taylor([EMAIL PROTECTED]): It appears that the default behavior of debian systems is to print Segmentation Fault rather than produce a core file. Can anyone tell me how to change this behavior for debugging purposes? Appreciate any help! Jeremy From man ulimit BASH_BUILTINS(1) BASH_BUILTINS(1) ... ulimit [-SHacdflmnpstuv [limit]] . -c The maximum size of core files created Seems to be a bash default. Yep. It may be useful to add that you may use the value `unlimited', as in `ulimit -c unlimited', which is probably what you want if you are debugging. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: BE MORE SIMPLE!!!!
I just did a quick search on the debian web site with the keyword HOWTO, this is what I got: Search Results Release Quality Package (size) stable100% doc-linux-es 1998.08-1 (739.8k) Linux documents in Spanish. stable100% doc-linux-fr 1999.01-1 (3077.4k) Linux docs in french : HOWTO, MetaFAQ ... stable100% doc-linux-it 98.05-1 (560.8k) Linux documents in Italian. stable100% doc-linux-pl 1998.03.29-1 (825.6k) Linux docs in Polish: HOWTO - ascii version stable100% doc-linux-pl-html 1998.03.29-1 (834.8k) Linux docs in Polish: HOWTO - html version stable100% lg-issue12 2-4 (219k) Issue 12 of the Linux Gazette. stable100% xcin 2.3.03-3 (1514.5k) Chinese input server for Crxvt in X11. Responses 1-7 of 7 responses shown. What puzzles me is why this list doesn't include 'doc-linux-text' or 'doc-linux-html'. Any guesses? In the first slink release, there was just `doc-linux', and `doc-linux-text' and `doc-linux-html' were not there. Maybe some search database needs rebuilding? Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: BE MORE SIMPLE!!!!
Debian installation manual is not at all the worst one. It is quite good actually. But it definetely isn't easy to find on the website instructions how to get started... This is getting on my nerves... * go to www.debian.org * note the section `Getting Started', which is the second below `What is Debian' * actually _read_ this section (all of both sentences), and discover you need to click on the link Release information in this section. * click * Note the header `New Installations', and the links Install Manual for SPARC Install Manual for Intel x86 Install Manual for Alpha Install Manual for Motorola 680x0 below it. * click on, say `Install Manual for Intel x86' (if that is what you want) * you now find yourself reading the installation manual All this involves *two* clicks from the main debian pages, and reading maybe 30 lines of text. If this `definitely not easy' for someone, I figure this person needs to acquire some more basic computer skills before attempting to install anything at all on their computer. Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: BE MORE SIMPLE!!!!
I'm the guy that asked which package had the HOWTOs. I know it seemed like a dumb question, but I did try searching on HOWTO in both dselect and on the Debian web site, and came up with nothing. And I started paging through the 2700+ packages shown in dselect, but gave up after an hour or so of that. Maybe it would help if the package was named doc-howto or something useful like that, or if it was installed as part of the Complete Developer Workstation profile I selected when installing Debian. The howto's are in doc-linux-text (as text) or in doc-linux-html (as ... html). I have a fresh slink install where they were in doc-linux. Maybe because of this transition they were hard to find? If you want to browse through packages, a good place to start is http://www.nl.debian.org/distrib/packages. There you can find short descriptions in several sections. I agree that it would be helpful for more questions on this list to be answered with instructions on how to find the information. But I see too many messages on this list (and throughout the Linux community) saying people are stupid or people are lazy when they ask questions that have obvious answers. People are often lazy. A lot of things might be improved about the debian website, but if people argue things like `it is hard to find the installation manual on the web site', they just didn't try. Most of the mails on this list tend to be helpful and informative, by any usenet or mailing list standard. The kind of reaction invoked by a question depends mostly on the tone. `I CAN'T FIND SHIT ON YER STINKIN WEB PAGE, NOW WHERE ARE THE HOWTOZZ?' is usually not responded very friendly to, whereas `How do I setup sound on my PC?' does get informative responses most of the time. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: [Installing Bo Debian on a 386 2 meg RAM?]
I am sorry to send this message again if you have already received it. It already appeared on the list. Hi, I have tried to install the Debian Bo distribution on my 386 with only 2 megs of RAM without success. Some people told me that they have been able to install it even if the machine does not have the 4 meg of RAM required. I have not been able to boot because the installation stopped saying that 4 meg of RAM is necessary. Please explain me what boot parameters must be given to overcome the 4 meg requirement. Is it possible to put swap on before booting? It is not possible to switch on swap before booting. I think people did not respond to your question because running linux in 2 MB seems very hard to do (if at all possible) and the result will not be very useful. If you really want to try it you should compile a _very_ minimal kernel for your system, and put that on the debian rescue diskette (which starts the installation). Maybe you should look around on the web if there are distributions specially geared towards such low memory systems. If you actually want to do anything with the system, I would advise you to try and find some extra memory. If you have 6 or 8 MB, you should be able to run a number of applications without X-windows nicely. With 8 MB it is possible to run X-windows (I did it, some years ago), but don't try to start anything like netscape. Good luck Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: OFFTOPIC: SDRAM PC100
On Wed, 2 Jun 1999 [EMAIL PROTECTED] wrote: The question is: if I buy a SDRAM PC100 chip, will it work on a computer with bus speed of 66Mhz ? Nope. I bought 2 128MB PC100 for my 66MHz ASUS, hoping that the dimms would still be usable when I get a new motherboard someday. I ran memtest86 for about 24 hours, no problems, and have been running fine ever since (several months). I started having problems recently (filesystem errors and signal 11) so I ran memtest86 again and one of the dimms has gone bad. I took it out and put back in an old regular 32MB dimm, and the two mixed togeter (160MB) ran memtest86 fine overnight, and seems OK after about a week. I don't know why the dimm went bad. I wouldn't think it's because the 100/66MHz issue. I would surprise me too. I also run PC100 memory at 66 MHz with future upgrades in mind. Please explain why you think PC100 shouldn't be in a 66MHz board. Do you have any ponters to info on the web? One thing that may be important to watch out for is that some older motherboards need physically different dimms. We have a dual PPro with an Intel mobo, which apparently needed special (read: rare) dimms with the `nudges' in different places than where they are in the current dimms. The current dimms simply don't fit in the mobo. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: c++ docs
On Mon, 31 May 1999, Brad wrote: Where are the docs for the c++ libraries? More specifically, i have a copy of C++ How To Program second edition here. It claims that #include sstream will allow strings to be manipulated as streams. sstream: No such file or directory g++ tells me. The next major release of g++ will probably have it. g++ currently has the older strstream, which descibes stream classes that operate on char*, not string. They can probably be used for mostly the same thing. The iostream library (including strstreams) is documented in the iostream info file, that comes with libstdc++2.9-dev. Just type `info iostream'. In the future strstream (which is not in the ANSI standard) will be replaced by sstream (which is in the ANSI standard), though I guess strstream will stay around for some time. Ok then, i'll just check the docs i think to myself. Tried the manpages. Tried info. Tried looking in /usr/doc. i couldn't even find anything about the c++ string class (which i know i have), much less using them as As someone else already pointed out, there is the stl-manual package, which contains the information from Silicon Graphics about the STL (g++ currently uses the SGI implementation). streams! Hmmm... did i miss a -doc somewhere?. Fired up dselect, and couldn't find any c++ docs at all, installed or uninstalled... So, does anyone know where the docs are? Or at least how to use something like what the book is talking about? Unfortunately, there is very little distributed with egcs nor in the way of actual documentation for either the compiler or the libraries. In the egcs-docs package there is extensive information about the compiler in info format. The description of the C++ libraries could be better, but the iostream library is descibed to some degree in the iostream info file. The C library is not part of egcs, but it is well documented in info format, and available in the glibcdoc package. It is actually readable (not just as a reference) and I recommend it to anyone who is programming in unix in C or C++. I expect information about the stl to pop up when g++ starts using the stl implementation from the egcs team. For now g++ uses the SGI stl and you can use its documentation. [... skip interesting references ...] HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Diamond Stealth 3D 3000, 4MB
I am not being able to make a decision about the xserver for the Diamond Stealth 3D 3000, 4MB. I have tried both xserver-svga and xserver-s3v. With the xserver-s3 I am able to use depth 32 in the mode 1152x864, but the system locks almost always when I close a xsession. I used to have this problem, and posted it to debian-user ages ago. Recently someone asked me if I had solved it already. I said I didn't, did some suggestion about ending the session with ctrl-alt-bksp instead, and then he came up with a solution that did work: From: Roberto Jung Drebes [EMAIL PROTECTED] X-Sender: [EMAIL PROTECTED] To: E.L. Meijer (Eric) [EMAIL PROTECTED] Subject: Solved (was: Re: Diamond Stealth 3D 3000 problems) In-Reply-To: [EMAIL PROTECTED] Message-ID: [EMAIL PROTECTED] MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Fri, 5 Mar 1999, E.L. Meijer (Eric) wrote: end an X-session with ctrl-alt-backspace? I think this sends a SIGTERM instead of SIGHUP. I cannot try this now myself, the card is in some I decided RTFM. If I edit /etc/X11/xdm/xdm-config and put the line DisplayManager*resetSignal: 15 Things appear to work.. Cool. This may work for you too. HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: compiler in PATH ?
On 18-May-99 David Z. Maze wrote: gcc isn't a C++ compiler, it's a C compiler. You should have g++ somewhere in your $PATH (if you install the Debian g++ package, in /usr/bin). If you haven't yet, install the g++ package and its dependencies. I realized later that gcc is not for C++ so I looked for any g++ compilers in my system and I found them: ii libg++2.8.2 2.91.61-1 The GNU C++ extension library - runtime vers ii libg++272.7.2.1-14.4 The GNU C++ libraries (ELF version). ii libg++272 2.7.2.8-0.1The GNU C++ libraries (libc6 version). ii libg++272-dev 2.7.2.8-0.1The GNU C++ libraries (libc6 version). The libg++2.8.2 package contains an old gnu C++ library that is no longer actively maintained. The 27 and 272 packages provide support for older versions of g++. You don't want to use these for compiling programs unless the program depends on the bugs in the 2.7.2 compiler. For slink, you need to get the following packages: g++ libstdc++2.9-dev libstdc++2.9 (you probably have that already) and the gcc packages. HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: FSCK!!! only 8 disks!!!!
WTF?!?!?!?!? Please someone tell me that linux isn't limited to 8 scsi disks! If it isn't tell me how the FSCK to fix this. I'm trying to set up a raid array with 10 disks after a couple crashes I read man MAKEDEV and found out that there is indeed a limit of 8 scsi disks. 8-( $ zless /usr/doc/HOWTO/SCSI-HOWTO.gz ... read read ... ./MAKEDEV sd\* should create entries for all SCSI disk devices (doing this should create /dev/sda through /dev/sdp, with fifteen partition entries for each) ... read read some more ... I say should because this is the standard unix behavior - the MAKEDEV script in your installation may not conform to this behavior, or may have restricted the number of devices it will create. If MAKEDEV won't do the right magic for you, you'll have to create the device entries by hand with the mknod command. The block/character type, major, and minor numbers are specified for the various SCSI devices in section ``Device Files'' in the appropriate section. Take those numbers, and use (as root) mknod /dev/device b|c major minor ... yup, that seems to be it. Boy, ain't that /usr/doc directory great :) HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: checkdir error
I have a shell script that unzips multiple files in a directory. This directory have 4 subdirectories named 001, ..., 004. Each one have 400 subdirectories named 001, ..., 400. In each of these 400 subdirectories I have 400 html files. The file system( using df ) still have enough space for the unpacking and, when running the 4th directory it prints anerror like this: checkdir error: cannot create 004/022 unable to process 004/022/04022001.html After that, everything that I try to do, like mkdir, etc. does not work.It says that there's not enough space on the device. You have probably run out of inodes. Each file or directory needs one. You are trying to create 4*400*400=64 files. You can check how many inodes are left using df -i By default there are 4096 bytes per inode, which means that for the amount of files you want to store your partition needs to be at least 64*4/1024=25000 MB in size (neglecting directories). If you really need to put the files in this partition, you need to reformat it and use the -i switch to mkfs.ext2. See man mkfs.ext2. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: checkdir errory
I wrote: You are trying to create 4*400*400=64 files. You can check how many inodes are left using [...] By default there are 4096 bytes per inode, which means that for the amount of files you want to store your partition needs to be at least 64*4/1024=25000 MB in size (neglecting directories). ^ Um, well, that should be 2500 MB. :) Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Can't allocate DMA buffer
Hi, I'm running the most recent kernel (2.2.7), and I also use a Crystal semiconductor sound card, so I have the cs4232 kernel module. Sometimes I get the kernel error: kernel: Sound error: Couldn't allocate DMA buffer I have plenty of available memory left, so I don't understand why I am getting this error. Has anybody else received this error? Does anybody know of a workaround or fix for it? I haven't received this error before installing the 2.2.7 kernel. I have had this error with kernels from the 2.0.x series. My floppy drive also has this problem sometimes. The problem is that apparently the DMA buffer needs to be allocated in the first 16MB of address space. In my view, the driver should reserve enough of this memory when it is loaded, but this may be a problem if it is not loaded at boot time (but later, as a module). I once asked around on a linux kernel news group, and people were thinking about how to reorganise the DMA memory. It doesn't seem to be trivial to solve. A `solution' would be to get a PCI sound card. Often you get around the problem stopping some applications hoping to free some of the `right' memory :( HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: finding and using applications
Suppose you have a Debian Gnu/Linux system set up and fully loaded with applications. A new user appears who is going to use the system. The new user is a unix novice. He/she knows enough basic commands to get by. Is there a simple way for that user to find every available application on the system, what the application does, and how to use it? To get an idea of what could be available, the debian web pages provide nice descriptions of the packages. To see what is installed on your system, just type $ dpkg -l | less Now if some short description seems interesting, try $ dpkg --print-avail package name Next action would be to see what is in /usr/doc/package name, and to try man and info pages. None of this seems very hard to me. I agree it should be advertised better. [...] A very concise way of finding out what commands are available to you in bash, is typing TAB twice, and then `y'. :) HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Advanced Printer Control?
I'm configuring a print server for remote clients. I've set up a Debian machine with an HP LaserJet 4000 hooked to it via parallel cable. I've installed lprng and samba, and can print to it from two types of client: 1) From Linux using BSD-style lpr. Text (notwithstanding line feed issues) and postscript print fine. I am looking at installing magicfilter. 2) From NT using local drivers, via Samba. Works fantastic. The issue, is that the NT drivers are able to take full advantage of the printer's capabilities. It can do duplex, 2- 4- 8-up printing, etc. How can I achieve such control from the Linux clients? I need to be able to specify at least duplex, and hopefully a greater subset of the printer's capabilities. Linux is working great as a print server to NT clients, but not so great to Linux clients. You can do two things about this: 1) Find out what control codes to send to your printer that will enable/disable duplex 2-, 4- or 8- up printing, and write a filter that sends those codes. 2) Use software. duplex: lpr has an option for muliple copies $ lpr -#3 bla.ps will get you three copies of bla.ps. Note that this is often disabled by default in /etc/printcap, and you will have to enable it there. n-up printing: Plain text files are best printed with enscript $ enscript bla.txt Enscript is able to print to rotated pages on one sheet of paper: $ enscript -2r bla.txt For postscript files, you can use psnup: $ psnup -12 bla.ps | lpr will print 12 pages per sheet on the default postscript printer. More than 2-up with plain text: $ enscript -p- bla.txt | psnup -4 | lpr Of course you can hide all the nice unix geekery to users in scripts. Single problem: psnup only works with `well-behaved' postscript files, that follow Adobe's document structure definitions (or what'sit-called). Any advice would be appreciated. I'm not even sure at which level (client? server?) this would occur. Pointers to HOWTOs that deal with this, and other resources, would be appreciated. PS: I can't help but imagine C++ iostream manipulators: cout unix_line_feeds file.txt duplex 4up file.ps; The last example is quite like that, except that it is written in the other direction. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: adding new libs (.a) to debian (gnome xml)
I tried to add the gnome_xml libs to debian. Due to the fact that i couldnt find a .deb I compiled the package, did make install which copied the said libs to /usr/local/libs. I then added the directory to /etc/ld.so.conf (i hope i'm not getting the file name wrong) and then ran ldconfig. In the subject you talked about .a files. These are static libraries. The ldconfig program is for configuration of dynamic libraries, which have names ending in .so. Then when i tried to compile another package which tried to link against the xml libs (gnome_print) i got an answear somthing like: error with -lxml no such library (I don't remember the exactly how it was writen). Probably because it tries to find the dynamic library. If anyone know of .deb instead of these two packages it would be better. I don't know of any debs, but before they arrive you could 1) Try to generate dynamic gnome_xml libraries (maybe you just need a different target from the makefile), install them, and then run ldconfig. 2) Specify the complete path to the .a file to link it statically, so gcc /usr/local/libs/libxml.a instead of gcc -lxml HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: xterm/top horizontal scrollbar
I have Top loading in an Xterm window in Fvwm2. I set Top to display the command line instead of the command name and some of the command lines are well off the right edge of the Xterm window. I can resize the window to view these command lines but I was hoping I could find an option to Xterm to add a horizontal scrollbar. I've looked at the man pages for Top and Xterm and don't see anything like that. I notice Xconsole has such a scrollbar. I was wondering if anyone knows how to set up Xterm with such a scrollbar or is there another option that would suit my wants better? Thanks, kent Your window manager may be able to do this. I am using fvwm2, and in the window menu there is an option `scrollbar' which enables a horizontal and a vertical scrollbar added around the window by the window manager, but that the xterm itself is unaware of. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Getting Rid of Unnecessary Files
[... please don't use mime on this list ...] I want to know if there is a way to clean out all of the copyright,copyleft,changelog, and other non-essential files in my Debian installation. i also want to clear out all of the non-English related stuff and the non-i386 stuff fron the tree. All of this takes up tons of space. These files are fine for learning purposes and for being able to establish contact with the various package maintainers. --however I have need of the disk space being taken up by these files. I have tried to remove them manually once before and many of them actually cause dependency faults if they are removed. Is it possible to remove them from a complete distribution tree of .deb files and repackage the distribution as a totally customized distro strictly for my own personal use? Did you try instead of removing the files making them 0 bytes? So for example if you have a list of files you don't need, you could do something like #! /bin/sh exec file_list while read f; do cat /dev/null $f done Then the files are still there, but take much less space. This may be enough to keep dpkg happy. On a more general solution, for reasons of disk space saving, a switch could be added to dpkg (and dselect and apt) not to put any of the files in the /usr/doc hierarchy on the disk. You could file this as a wish-list item in the bug system if it is important to you. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
ddd's segfaulting tradition
[discussion about ddd problems...] I wonder, does anyone use ddd in a serious way with C++? Everytime a new debian release arrives I give it a try, and everytime it manages to segfault within a few minutes. I suppose ddd should be nice for C++ if it worked, but I never found one real life bug with it in _my_ code before I hit one in ddd itself. Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: SHELL environment variable.
I was always under the impression that the SHELL envirinment variable was supposed to point to the _current_ shell. In my setup at least, it doesn't. It _always_ points to the login shell (I change to a different shell but the variable remains the same). Bug, feature, user error? Any comments or suggestions welcome. It points to the login shell because it is the login program that sets it. The shells don't do anything with it, they just inherit it from their parent process. See man login. If you want to determine what shell you run, you may be luckier trying to read $0, although that gives rather erratic results as well (e.g. bash gives adds a `-' in front if it is a login shell, csh doesn't seem to recognize this at all). HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: oddness from getgrnam()
I was testing a program and to see how it handled an invalid group I did:
gr = getgrnam(bob);
Now obviously this failed. However the string from perror() states:
Could not find file or directory
Why is this? Seems like a counter intuitive error.
I think the error is not set by getgrnam(). If I run
#include stdio.h
#include stdlib.h
#include grp.h
int main(void)
{
struct group* gr;
perror(before);
gr = getgrnam(tgakem);
perror(after);
return 0;
}
it always prints
before: Success
after: No such file or directory
It makes no difference if tgakem exists (which it happens to do) or
if it doesn't. My guess is that getgrnam tries to look for group
descriptions in several files, of which at least one does not exist.
If I run the above program with strace (strace ./prg) and grep for
`open', I get
open(/etc/ld.so.preload, O_RDONLY)= -1 ENOENT (No such file or directory)
open(/etc/ld.so.cache, O_RDONLY) = 3
open(/lib/libc.so.6, O_RDONLY)= 3
open(/etc/nsswitch.conf, O_RDONLY)= 3
open(/etc/ld.so.cache, O_RDONLY) = 3
open(/lib/libnss_compat.so.1, O_RDONLY) = 3
open(/lib/libnsl.so.1, O_RDONLY) = 3
open(/lib/libnss_files.so.1, O_RDONLY) = 3
open(/etc/group, O_RDONLY)= 3
As you can see the first file could not be opened. I checked and it
indeed does not exist on my system, so that will have triggered the
error message. The getgrnam function returns NULL on failure, and
apparently does not manipulate errno.
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: AGP video for Linux?
Personally, I don't even know what AGP video is, so I am forwarding this to the debian-user mailing list. Bob Erik Olsen [EMAIL PROTECTED] writes: I was wondering if AGP video is supported by Linux yet. I did not see it on the list of compatable devices/bus archetectures but was just curious. AGP is a variation on the PCI bus specially for video cards. If you want to know if a video card is supported, you have to check out the XFree86 compatibility. Usually if a certain card or chip set is supported, the AGP version is supported, so you can use it in X. If you are in doubt about a certain card, just ask around on the various mailing lists and news groups. If you want to use a card right now, it is usually a bad idea to pick one that is only going to be supported `real soon now'. Another matter is support of the specific feature that AGP was meant for: storing textures in main memory and sending them to the video card at high transfer rates. I don't know if there is any support for that. HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: mount partition or disk
Hello, is there an easy (or canonical) way of (auto)detecting whether a disk is partitioned, please? What I have is a SCSI magneto-optical disk drive (230MB removable disks) and a bunch of disks. With some of the disks, I need to mount /dev/sda1, with others /dev/sda directly. (The disks have existing data on them, written under DOS, obviously various versions of either DOS or the driver.) Any suggestions? You can do this with `fdisk -l /dev/sda'. You either have to be a member of group disk, or be root, or mess with the default permissions (not advisable) to issue this command. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: X thinks my screen is larger than it actually is
Thanks for this response, but it doesn't apply in my case. I don't have a virtual screen set up, the only resolution mention in my file is 640x480. The laptop pages don't help - frankly because it isn't an issue with laptops - it's an X server problem. I get the same results using a desktop. The Xservers (or window managers) simply don't stay within the confines of 640x480 resolution - and they should. If I maximize a window, it puts it in the full 640x480 window - so why does X/window-managers put these windows off the screen when the are first painted? That's just sloppy.. I know in most cases I can reposition them, but I shouldn't have to. If I only have 480 pixels high, why does a window pop up that goes beyond this limit? -Jay This is not sloppy of X/window-managers, this is sloppy of the client programs you are running. Programs decide how large they want their windows to be. Many X11 programs can be configured using X resources to have a certain default size. To learn about resource specifications, have a look at the man pages of X and xrdb, and at the files in /usr/lib/X11/app-defaults/. A common way to handle your personal preferences, is to have a file .Xresources in your home directory, and read it with the xrdb command in your .xsession file, like xrdb -merge $HOME/.Xresources The size of windows is determined with the geometry resource. E.g., in my .Xresources I have the following lines for xdvi: XDvi*shrinkFactor: 4 XDvi*expert:1 XDvi*sideMargin:1.5 XDvi*background:NavajoWhite XDvi*geometry: 1266x993+0+0 This asks for a window size of 1266x993 pixels. Note that this is without the window borders, since the xdvi program cannot know what borders the window manager puts around them. Besides all this, there are X programs that simply assume that you have a higher resolution than 640x480. This is because many X programs originate from unix workstation platforms, which rare have resolutions lower than 1024x678. HTH, Eric Meijer -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: [off-topic] How to auto-run xlock?
I have a program that I want to run only while my PC is unattended. I
don't need the screen locked necessarily.
xlock does this for me:
xlock +nolock -startCmd startsetiathome -mode blank
but I have to run it manually. When I log back in, it does kill the
running command (startsetiathome) just like I want.
I'd like this exact behaviour but somehow I need xlock to always be
running, waiting to jump in after the idle period.
I thought about this problem before, and now realised you can check
keyboard and mouse activity looking at /proc/interrupts. You could then
start a script like this in your .xsession file (adapt the variables at
the start to your needs):
runsaver:
#! /bin/sh
# these are the interrupts of your keyboard and mouse,
# check out your /proc/interrupts:
INTERRUPTS=1 4
# this is the interval in seconds between checks of /proc/interrupts
SLEEPTIME=30
# if more than MAXINTERVALS * SLEEPTIME seconds pass without activity,
# the xlock program will run
MAXINTERVALS=10
# command line options for xlock
XLOCKFLAGS=-nolock -mode bat
checkactive () {
TOTALS=
INTERVALS=0
while true; do
OLDTOTALS=$TOTALS
TOTALS=
for x in $INTERRUPTS; do
TOTALS=$TOTALS `grep \$x: /proc/interrupts | \
sed -e 's/^.*: *\([0-9]*\).*/\1/'`
done
if [ $TOTALS = $OLDTOTALS ]; then
INTERVALS=`expr $INTERVALS + 1`
if [ $INTERVALS -ge $MAXINTERVALS ]; then
break
fi
else
INTERVALS=0
fi
sleep $SLEEPTIME
done
}
while true; do
checkactive
xlock $XLOCKFLAGS
done
Note I have only just barely tested this.
If anyone is interested, the program that I'm running is the SETI at Home
client. SETI is The Search for Extraterrestrial Intelligence.
Great, sent my regards to the aliens once you find them :)
HTH,
Eric
--
E.L. Meijer ([EMAIL PROTECTED])
Eindhoven Univ. of Technology
Lab. for Catalysis and Inorg. Chem. (SKA)
Re: vi in Debian (slink)
However, vim is not standard. I routinely work on HP-UX these days and doubt that vim is installed there, for example. It usually is not too hard to go to a debian site, download the original source tarbal, and compile it for personal use. I just did that with procmail on SGI. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
elm filter?
Hi all, I am receiving mail on a SGI system, that has no procmail, but the elm filter program. I would like to use the same thing on debian, but I cannot seem to find it. Is it anywhere available in slink? Thanks, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: How to open xterm with and a program at once
Hi, how do I open xterm with a program? Let's say I want an icon on my desktop opening mc or mutt. What would the command be to have an xterm with one of these programs open automatically with xterm? xterm -e program This, by the way, is in the man page of xterm. If you'd type `man xterm', and then `/execute', you'd find yourself in the middle of the description of this option. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Mount for normal user
How can I mount devices (CDrom / HD) as a normal user. For example : $ mount /dev/hdb /mnt/hdb mount: only root can mount /dev/hdb on /mnt/hdb Normally the mount command checks if you really are root, and disregards group membership. You can allow users to mount a certain partition by adding the `user' option in /etc/fstab. For example, I have a line for a cdrom in my /etc/fstab that looks like this: /dev/hdd/cdrom iso9660 defaults,ro,user,noauto 0 0 The user option allows every user to mount and umount the cdrom, and noauto prevents the system from trying to mount the cdrom at boot time, which can be annoying if there is no cdrom in the drive. I've tried to add my user (sami) to some groups, but... : sys:x:3:sami adm:x:4:sami disk:x:6:sami lp:x:7:lp,sami mail:x:8:sami voice:x:22:sami cdrom:x:24:sami floppy:x:25:sami sudo:x:27:sami audio:x:29:mary,sami majordom:x:31:majordom Where is any help about Debian's groups... ? I don't know. Most groups are self-explanatory. However, How to remove user from groups (without VI) and is there a way to have all root permission without being root (UID 0) because some programs don't want to run as root ? If programs do not want to run as root there usually is a security reason for that. This means it can be in some way dangerous to your system, and you'd better not try to run it as root anyway. You can remove users from a group using the usermod command and the -G option. With the -G option you list the additional groups (beside the default group) that a user is a member from. If the user was listed with other groups as well, (s)he is removed from those. Check out the manual page. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
Re: Slink upgrade and /usr/local problems locating libs!
Hey guys, Since I have upgraded to slink I have noticed that when I unzip certain tars in /usr/local the binaries seem to have trouble finding the required libs. For example, Wingz3 can't find libXpm.so.4 although it is in /usr/X11R6/lib. Wingz3 worked with hamm. Has there been some kind of change in Slink that would cause this? Is the library really there, or is there just a dangling symbolic link? On my slink system /usr/X11R6/lib/libXpm.so.4 is a symlink to libXpm.so.4.10. If the latter file is not there, you need to install the xpm4g package if the Wingz3 is a libc6 program. If Wingz3 is a libc5 program, you need to install the xpm4.7 package from the oldlibs section. Check this out with `ldd Wingz3'. HTH, Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
tar oddness
Hi all, I was pondering on the problem of splitting output over multiple mountable media, and discovered the following: $ tar czf foo elmfract $ tar czf - elmfract bar $ ls -l foo bar -rw-r--r-- 1 tgakem users 10240 Mar 29 16:59 bar -rw-r--r-- 1 tgakem users7950 Mar 29 16:59 foo The file bar is a valid archive, but contains trailing garbage. Any idea how this comes about? It does not happen if the z flag to tar is not present. I noticed this when I wrote a small program that takes input from stdin (through a pipe), and writes this to different files on different volumes of a mountable medium (say, floppies). Eric -- E.L. Meijer ([EMAIL PROTECTED]) Eindhoven Univ. of Technology Lab. for Catalysis and Inorg. Chem. (SKA)
