Re: [expert] OT: Databases
On 03.18 Pupeno wrote: On Monday 19 March 2001 14:11, you wrote: I think MySQL (included with MDK 7.2, probably earlier releases also) would fit your needs, running on the local workstation. It is open source, free, and (from what I have been told) has pretty good performance. I cannot tell people that, to use my small, simple, little program they have to configur a mysql, I'm programming for users, not administrators. I would take a look at the standard and simple Berkely DB. It was part of glibc, but now there are three versions in mdk, db1, db2 and db3. I think the latest release is db3, and the other versions are included for compatibility. That is what rom4 uses for the rpm database. Install db3, db3-devel, and db3-utils and look into /usr/share/doc/db3-devel-3.1.17 for the api description (both C and C++). Hmmm, now that I think about it, you will have db1-2-3 in Cooker, in standard 7.2 you probably have only just the old db. But you can always get the src.rpm and rebuild. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac20 #2 SMP Tue Mar 13 15:49:10 CET 2001 i686
Re: [expert] Helix-Gnome supports Mandrake! -- Official Word
On 03.19 Benjamin Sher wrote: There are some bugs in helix-update that prevent us from releasing the most critical package (an updated gdm) through normal channels, so at this point it looks like we're waiting for Red Carpet to mature a bit more. In the meantime, the updated gdm package can be downloaded from http://primates.ximian.com/~peter/gdm-2.0beta4-22mdk_helix_10.i586.rpm I understand that Ximian has done a great job with Gnome, but in Mdk8.0 all the gnome packages are more up to date that those of Ximian. For example, that updated package is gdm-2.0-beta4, and current cooker gdm is 2.2.0, that was recently released. By the time they support 8.0, there will be no need of Ximian in a Mdk system. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac20 #2 SMP Tue Mar 13 15:49:10 CET 2001 i686
Re: [expert] hi all
On 03.18 Clint Olson wrote: --snip-- anyways the new nvidia drivers are cool ! i've just installed them to my 7.2 mandrake box and it's stable and fast. ---snip I agree. I don't know what everybody means about NVIDIA's Linux drivers being binary-only. Personally, the (binary) RPM didn' work for me, and I had to download their source tarballs and compile them. After that they worked fine :) Look carefully to your .tar.gz 'source' driver. You will see three files: -rw-rw-r--1 nobody nogroup644928 Mar 13 15:33 Module-nvkernel -rw-rw-r--1 nobody nogroup280133 Mar 13 15:33 nvrmapi.lib -rw-rw-r--1 nobody nogroup265880 Mar 13 15:33 xf86nvrmapi.lib that are the big deal of the drivers. They are binary and they are what does the hard work. You do not have the source for those files. The rest of the files, what sure are C source files, are just glue code or wrappers between the kernel and nVidia's code. Which is the difference ? For example, the 767 version crashes with some systems and kernel AGP. If those .lib were in source format, someone would have yet found the cause. This way, you have to wait that nVidia people can reproduce the crash, trying to build a system similar to those that lock up. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac20 #2 SMP Tue Mar 13 15:49:10 CET 2001 i686
Re: [expert] Wget as a mirroring utility
On 02.26 Tom Eastman wrote: Hi there all, I'm currently using wget to mirror the Mandrake cooker ftp, but I was wondering... Wget has a lot of functions of a mirroring program, eg time-stamping and stuff, however, what it DOESN'T seem to do, is delete a local file that no longer exists on the mirrored ftp server, is there a way of telling it to do this? Or should I use one of the other mirroring tools... like mirror? :-) Does anyone have an already-set-up mirror config file for the cooker? Use lftp. I works fine for mirroring (timestamping, size comparison, mirror over ftp or http and ssl versions, and deletes local files not present in remote source). It is worth a look at. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac4 #2 SMP Mon Feb 26 00:21:23 CET 2001 i686
Re: [expert] Offtopic, sortof?: Apple Macs and networking
On 02.23 J . A . Magallon wrote: Then configure netatalk in the linux box to hang on the same AppleTalk network. Quoting to myself, there is a flag -zone "My zone" in /etc/atalk/atalkd.conf. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac1 #2 SMP Fri Feb 23 02:34:42 CET 2001 i686
Re: [expert] Offtopic, sortof?: Apple Macs and networking
On 02.22 Praedor Tempus wrote: I am connecting my laptop into my work/university network which is dominated by windoze and some unices. There are a scattering of Macs around, including those in the lab I work in. I do not have a "name" for my computer - it is whatever IP address I get assigned on a particular day (today I was 145.100.225.101 - the alternative name is something like ds101.hmbg.wsu.edu). First fo all, plz, in the next reply please tell me you MacOS version (8.x, 9, 9.04, 9.1...) You can always have an AppleShare name on your imac, that has nothing to do with IP names, so you can name your mac 'Preston' in the File Sharing control panel, and this is idependent from TCP/IP access, for which you mac is named ds101.hmbg.wsu.edu. For AppleTalk nets, your iMac will be that name. I am trying to get setup so I can connect/share with them from the iMac to my linux system rather than the normal linux netatalk server - Mac. I have netatalk installed and running and can connect from the iMac to my linux system via appleshare without difficulty - though my system is not shown when I open up the iMac's chooser and select "Appleshare". Instead, I have to select the IP option and then manually enter the IP address of my linux system, then I can mount my linux $HOME directory on the iMac. That is fine, as far as it goes, but I would much rather be able to go the other way, connect to the iMac from my linux system and mount the Mac drive on my linux system. Then something is not working well. When you open you Chooser and click on the AppleShare icon, what do you see ? Are there many zones available or just one ? Open the AppleShare control panel in the imac and you should see which AppleTalk network zone your iMac is hung on (and this is also the default zone selected when you click on Chooser-AppleShare). Then configure netatalk in the linux box to hang on the same AppleTalk network. So, looking at the setup, it appears that the iMac is all set to permit file sharing, and indeed, if I boot to windoze, I can access the iMac. How ? Which protocol ? Perhaps someone who has mixed Mac and linux networks at home or wherever could give afpfs a shot? I got it at http://www.panix.com/~dfoster/afpfs/ It DOES compile and "run" on 2.2.x kernels. Don't know about 2.4.x kernels. I am trying to make it run on 2.4, but I think I do not have enough kernel knowledge to do that... -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac1 #2 SMP Fri Feb 23 02:34:42 CET 2001 i686
Re: [expert] missing hd?
On 02.24 Tom Strickland wrote: On Fri, Feb 23, 2001 at 11:43:27AM -0800, Ron Heron wrote: hmmm, are you sure it's on hdd? hdd is the slave of the first ide slot. AFAIK, hdd is the slave in the SECOND ide channel... If fdisk says no response, then it is truly missing! how many beers did you have when you did this? :) Ah - if only I could ascribe this to beer! Too much coffee perhaps... anyway, as I've already stated, if I restart from the installer cd-rom and use diskdrake from there as part of the installation procedure, I can see hdd with no problem. I even went Lets look at this (perhaps I have too much coffe also, and all that comes is stupid). If your disk is the slave in the first ide channel, it should be named hdb. So I assume you have one other hd and a cdrom. Lets suppose: Your hds are on the IDE 1: hda (the old) and hdb (as you should see it) Your cdrom is at hcd (master at IDE 2). And if your bios swaps your ide channels when booting from cd at IDE 2, and hdb becomes hdd ? -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.2-ac3 #1 SMP Fri Feb 23 21:48:09 CET 2001 i686
Re: [expert] rpm/deb files
On 02.03 Ted Wager wrote: Is there a way to view the contents of a .rpm and a .deb file as you can with a .tar file or a way to strip the .deb or rpm file down so as to see the contents without installing it ?? rpm -ql *.rpm, for an rpm. for a deb, i dunno... -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac2 #1 SMP Sat Feb 3 10:45:59 CET 2001 i686
Re: [expert] X and monitors - how to switch?
On 02.02 Neal Lippman wrote: I notice, for instance, that if you change the monitor on a Win system, you don't seem to need to worry - is that because Windows is better at PNP detecting the monitor and properly setting it up or because by default it uses such conservative settings that it is unlikely to exceed the monitor's specs? X also do that 'PNP detection' you talk about. If the monitor is DDC capable (most actual monitors are) X asks the monitor its valid screen modes (freq, res). It works since 4.0, I think. If windows could do that, vendors would not distribute that floppys with the monitor containing the 'monitor.inf' file... -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac1 #2 SMP Fri Feb 2 00:19:04 CET 2001 i686
Re: [expert] ssh reverse mapping problem
On 02.02 Bug Hunter wrote: This is an indication that reverse dns does not work for user from ip address aa.bb.cc. try adding that ip address in /etc/hosts. that might help. On Thu, 1 Feb 2001, Svante Signell wrote: Shall I take this seriously? Which programs to check? Where to look? Feb 1 20:20:10 computer sshd[864]: WARNING: /etc/ssh/primes does not exist, using old prime Feb 1 20:20:15 computer sshd[864]: reverse mapping checking getaddrinfo for aa.bb.cc failed - POSSIBLE BREAKIN ATTEMPT! Feb 1 20:20:15 compaq sshd[864]: Accepted password for user from xx.yy.zz port 1020 ssh2 If user is connecting via PPP or DHCP (cable), it is possible that his ISP has implemented DNS only in one direction (ip-name or name-ip). See if you can nslookup or host the name and the address. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac1 #2 SMP Fri Feb 2 00:19:04 CET 2001 i686
Re: [expert] FW: RPM annoyance
On 02.02 Lou Baccari wrote: vvv only packages with major numbers = 3 are supported by this version of RPM ^^ error: xfig-3.2.3c-3.i386.rpm cannot be installed rpm package is version 4. rpm-3.0.4 can not deal with v4 packages, 3.0.5 can. If you do not want to go to rpm-4, at least you must upgrade to 3.0.5. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac1 #2 SMP Fri Feb 2 00:19:04 CET 2001 i686
Re: [expert] lm sensors and i2c
On 02.02 PBone wrote: I am trying to get temperature monitoring for my athlon/epox 8kta2/Mandrake 7.2 system. I have been to the lm_sensors site but the latest Mandrake rpms are for Mandrake 7.0. Is there an easy (ie rpm) way to install lm_sensors- (I have had low success rate trying to compile programs under Manrake 7.2) cooker: kernel 2.4 + lm_sensors -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac1 #2 SMP Fri Feb 2 00:19:04 CET 2001 i686
Re: [expert] FW: RPM annoyance
On 02.03 Altoine B. wrote: "J . A . Magallon" wrote: On 02.02 Lou Baccari wrote: vvv only packages with major numbers = 3 are supported by this version of RPM ^^ error: xfig-3.2.3c-3.i386.rpm cannot be installed rpm package is version 4. rpm-3.0.4 can not deal with v4 packages, 3.0.5 can. Correction, the package that he is trying to install is rpm3 or less while his package manager is rpm 4. That particular package would only unpackage with a package manager of rpm 3.0.5 or less. I recommend getting the source for that package and compiling it that way. That is how you understand this ??: only packages with major numbers = 3 are supported by this version of RPM ^^^ so this rpm is only for = v3 packages, and package is v4. AH, i got you, when I said 'rpm package', I did not meant rpm.rpm, but xfig.rpm (as opposite to, for example, deb package or tgz package...). -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-ac1 #2 SMP Fri Feb 2 00:19:04 CET 2001 i686
Re: [expert] An observation on this Mandrake list
On 01.31 Simon Cousins wrote: What I'm seeing more and more, however, and what makes me a little sad and nostalgic for the pioneering days, is a general movement towards "complaining to the developers" 'cause something doesn't work to a user's full expectations. I don't think so. Apart from an small amount of people who see the list as a free support place, I see the usual 'complaint' messages as: hey, this does not work in my config, so check it out to clean up your marvelous work. Only in few cases you can see the hidden added phrase: 'and have it ready for me this night, come on, the only thing you work for is to clean my bug'. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-zcn #1 SMP Tue Jan 30 11:38:19 CET 2001 i686
Re: [expert] rh6.1 to mandrake 7.2 makefile porting
On 01.29 Stefano Perticoni wrote: Hi Assume makefiles are included by another in the main() dir. Also assume the RedHat makefile is correct couse it compiles. Perhaps it works (compiles), but you are using things that are messing up the compilation process and can break your Makefile even if you update the compiler. It works by chance. Why the RedHat works while the Mandrake doesn't ? See above. LDFLAGS = /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.91.66/libg2c.a \ Never do this. It is supposed that gcc is smart enough to know where to look for its own libraries. So delete that .a. $(AZTEC_DIR)/lib/libaztec.a AZ_LIBS = -L$(AZTEC_DIR)/lib -laztec -lm This is what LDFLAGS ^ should contain instead. HASRANLIB = t LIFEV_DIR = $(HOME)/fem_Life_5/Software SPARSELIB_DIR = $(LIFEV_DIR)/lib/SparseLib++/sp1_5c IML_DIR = $(LIFEV_DIR)/lib/IML++ AZTEC_DIR = $(HOME)/aztec LIB_CPP1 = /usr/include/g++/ LIB_CPP2 = /usr/lib/gcc-lib/i686-pc-linux-gnu/pgcc-2.91.66/include LIB_CPP3 = /usr/include/g++-2/ INCLUDES= -I$(LIFEV_DIR)/include/mesh \ -I$(LIFEV_DIR)/include/fe \ -I$(LIFEV_DIR)/include/util \ -I/usr/include/g++ \ ^ not needed -I/usr/lib/gcc-lib/i686-pc-linux-gnu/pgcc-2.91.66/include \ ^ not needed -I$(SPARSELIB_DIR)/include \ -I$(SPARSELIB_DIR)/mv/include \ -I$(AZTEC_DIR)/lib \ -I$(IML_DIR) \ -I$(LIB_CPP1) \ ^ not needed -I$(LIB_CPP2) \ ^ not needed -I$(LIB_CPP3) ^ not needed You can't use includes from three diffrent versions of the standard c++ library. In fact, none of them are needed, because g++ knows where to look for them. And you should NEVER include things like /usr/lib/gcc-lib/i686-pc-linux-gnu, because you can mess a compiler by forcing it to take includes or libs from a path belonging to one other. Just do not put any of that unnecessary includes, and g++ in both RH and MDK will know where to look for things automatically. Let the compiler do its job. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-pre11 #1 SMP Mon Jan 29 01:18:32 CET 2001 i686
Re: [expert] Summary: How to move /usr to another partition
On 01.26 David Dennis wrote: why take a chance your links or dot files will be missed with a mv or cp when tar gets it all and preserves everything 100% with one command ? just curious. 'cp -a' does it all right, I used it to change my drive and cloned all my system. It works ok with devices, special files, permissions, links, etc. Just clones whatever you give. See man cp. (btw, much much faster that tar, cpio, etc) -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-pre10 #4 SMP Wed Jan 24 00:20:15 CET 2001 i686
Re: [expert] How dost one mount a Mac drive?
On 01.23 Praedor Tempus wrote: How do I connect to and mount a Macintosh harddrive that is networked AND has filesharing enabled (and tcpip enabled)? In short: YOU CAN'T (afaik). You want this scenario: macs are servers, share their files, and you want to see them from the linux box. Macs share their files through a protocol called AppleTalk, that works over the same physical ethernet than TCP/IP (ftp, telnet, nfs), but is not the same. So you want an AppleTalk client running in your linux box to connect to appletalk servers and mount their disks. There is no client to do that. The latest news I have heard about that are in: http://www-sccm.stanford.edu/Students/hargrove/HFS/index.html#Clients http://www.panix.com/~dfoster/afpfs/ so it seems to be in very unusable shape. The reverse (see your linux box from macs Chooser), can be done with 'netatalk'. Look at: http://netatalk.sourceforge.net/ And you can install an ftp server on your macs. NetPresenz it the best for me. Get it in you favourite mac mirror, and put an alias in the Startup Items Folder. I do not know if it will be still free... BTW: anybody knows anything about linux support for - HFS+ - afpfs -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.1-pre10 #4 SMP Wed Jan 24 00:20:15 CET 2001 i686
Re: [expert]
On 2001.01.20 je gianopoulos wrote: subscribe expert Send this to 'majordomo@', not 'expert@.' -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac10 #1 SMP Sat Jan 20 10:43:18 CET 2001 i686
Re: [expert] Kernel 2.4 versus supermount
On 2001.01.19 Chris Spencer wrote: On Fri, 19 Jan 2001, Mogens [iso-8859-1] Jger wrote: Hey list. I am running a system with a dual Celleron, so I use the -smp- kernel. So far I can only use the 2.2.17.. because the pppd is NOT supported by the kernel, when I install the kernel-smp-2.4.0-5mdk.i586.rpm! why? will it be activated in a comming compile? So I downloaded the latest 'tar' of the kernel, and everything compiled fine, but I cant find the supermount option, so my cry for help is: how do I activate/include the supermount option in a "home-compile"? You need to download the diff and patch the kernel to get supermount to work. Off hand I don't remember what the web address is but if you go to freshmeat.net and do a search on supermount you can go to the author's web page, download the diff, and patch your kernel. You will probably have to change the diff file because by default it tries to patch files in a rather obscure directory. If you open up the diff the first three lines will show you the directories that the patch is going to look for the files. Do a find and replace and change the path to whatever your path to the source is. Once its patched you will find the option to enable supermount. http://www.geocities.com/alexismikhailov/supermount.html -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac9 #2 SMP Sun Jan 14 01:46:07 CET 2001 i686
Re: [expert] Problems with 2.4.0 and the reiserfs and supermount patches (and the mdk rpm)
On 2001.01.19 Jay Woods wrote: I had understood that the kernel wouldn't compile with gcc 2.96 but rather no later that 2.95 (according to Linus). ---Jay The only compiler that is granted to compile kernel 2.4 ok is egcs-1.1.2. It is guaranteed that it generates correct code. For 2.95.2, probably it makes no mistake, but IS NOT GUARANTEED. That means that most of the things will work, but perhaps in your particular setup you use a driver that is miscompiled by 2.95. And 2.96 is the same, its optimizer was horribly broken (I saw a test in kernel list that showed 3 lines of C code that when optimized-reorganized by gcc-2.96 produced INCORRECT assembler results like first using a variable and after this asigning the initial value). But the state of 2.96 is much much better now, so you can even trust a bit on it to compile a kernel. The fact is (as I have seen in the kernel list) that when 2.96 is guaranteed not to be buggy, will produce better code than egcs and 2.95. In mandrake case, the better setup is to have gcc-cpp, gcc, gcc-c++, libstdc++ and libstdc++-devel (and if you need it, g77 or objective c) from the 2.96 version. And it you do no trust 2.96 to build kernels, install egcs and egcs-cpp (just those two packages, no egcs-c++ nor any egcs-x more). egcs installs a binary called kgcc, that recent kernels can detect and use to build kernels. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac9 #2 SMP Sun Jan 14 01:46:07 CET 2001 i686
Re: [expert] RPM updates
On 2001.01.18 Andrew George wrote: The gotcha is (of course there is one) that RPM dosn't handle dependancy problems that well and you can run the command and just get a screenfull of broken dependancy links thrown back at you for your trouble Especially when package conventions have changed from 7.2 to cooker in the sense that a 7.2 package that was pack.x.y.z pack-devel.x.y.z now is split in libs and binaries and looks as: libpack.x.y.z libpack-devel.x.y.z pack.x.y.z So rpm -F only tries to update pack, and misses libpack- -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac9 #2 SMP Sun Jan 14 01:46:07 CET 2001 i686
Re: [expert] Problems with 2.4.0 and the reiserfs and supermount patches (and the mdk rpm)
On 2001.01.18 Christian A Strmmen [Number1/NumeroUno] wrote: I entered the 2.4.0 src dir and did "make mrproper ; make xconfig ; make dep ; make modules ; make modules_install ; make bzImage", however, during make modules I got this error: snap In file included from /usr/src/linux/include/linux/kernel.h:11, from rtc.c:60: /usr/lib/gcc-lib/i586-mandrake-linux/2.96/include/stdarg.h:43: parse error before `__gnuc_va_list' /usr/lib/gcc-lib/i586-mandrake-linux/2.96/include/stdarg.h:43: warning: data definition has no type or storage class /usr/lib/gcc-lib/i586-mandrake-linux/2.96/include/stdarg.h:110: parse error before `va_list' /usr/lib/gcc-lib/i586-mandrake-linux/2.96/include/stdarg.h:110: warning: data definition has no type or storage class In file included from rtc.c:60: /usr/src/linux/include/linux/kernel.h:62: parse error before `va_list' /usr/src/linux/include/linux/kernel.h:62: warning: function declaration isn't a prototype cpp: -lang-c: linker input file unused since linking not done It seems your cc install is very screwed, the -lang-c does not appear in any kernel Makefile, so I suppose it comes from the cc driver. It looks like it is calling /usr/bin/cpp for preprocessing (so it does not understand the -lang-c flag the cc command gives him, because it is only internal, for cpp0), and it should use the internal cpp0, usually located at /usr/lib/gcc-lib/i586-mandrake-linux/2.96/cpp0. As a suggestion (as you seem not to be using egcs to build kernel) check you don't have any egcs- package installed (specially egcs and egcs-cpp). You seem to be using cooker (at least gcc 2.6). So get the latest gcc-2.96 (gcc, gcc-cpp). Perhaps you miss the gcc-cpp package, and your gcc-2.96 is trying to use and old cpp. And if you want your life to be easier, get the 2.4.0 original sources from www.kernel.org, and the 2.4.1-pre8 patch, to run a 2.4.1-pre8 kernel. The advantage (apart from tons of bug fixes and cleanup) is that reiser is yet integrated in the main kernel tree for the 1-pre series from Linus. Then, you just have to look for a suitable supermount patch (I don't know where to get supermount 'cause I hate it). -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac9 #2 SMP Sun Jan 14 01:46:07 CET 2001 i686
Re: [expert] RPM updates
On 2001.01.19 Mark Weaver wrote: So then it's better to do the -Uvh instead of the -Fvh? When moving from 7.2 to Cooker, yes. Once all your system is a mess^H^H^H^HCooker, -F will be usefull again. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac9 #2 SMP Sun Jan 14 01:46:07 CET 2001 i686
Re: [expert] arrrrgh!!! (/usr/sbin/update-alternatives)
On 2001.01.14 Tib wrote: This is getting frustrating I tried to update the bzip2 rpm, and this is what I get: [root@unica seq]# rpm -Uvh bzip2-1.0.1-6mdk.i586.rpm error: failed dependencies: libbz2.so.0 is needed by ImageMagick-5.1.1-9mdk Now, I could not find a libbzip2 rpm at the mandrake RPM area at download.sourceforge.net/pub/mirror/mandrake/current/i586/Mandrake/RPMS, and the only one I could find at rpmfind.net was a cooker rpm and it gave me an error trying to install that one. NOW what do I do? You're sinkin into Cookeri like it... Try atik.ciril.fr. It usually goes fast for me. And fr.rpmfind.net. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac8 #1 SMP Fri Jan 12 18:02:50 CET 2001 i686
Re: [expert] Multiple video adapters
On 2001.01.09 Anthony Russello wrote: Has nyone ever set up multiple monitors under linux (using two separate video cards)? Would I be able to run a console on one screen, and an X session on another and be able to somehow switch between the two? Or will both monitors only be available inside X? If anyone has any information regarding this, I would greatly appreciate it. I could find almost nothing on xfree86.org in their documentation. XFree 4.0+ has the ability to do that. You have to define two "Device" sections, two "Monitor" sections, two "Screen" sections and setup the section "SeverLayout" for your setup, configuring things like 'screen 1 is right of screen2'. Usually they are independent screens, but also you can use 'Xinerama'. Look at: http://www.xfree86.org/4.0.2/RELNOTES4.html -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0-ac4 #1 SMP Mon Jan 8 22:10:06 CET 2001 i686
Re: [expert] Sounblaster PCI 128
On 2001.01.05 Richard Humphrey wrote: Someone just gave me a SoundBlaster PCI 128 Card and I was wondering if this card will work with Linux Mandrake? I am currently using v 7.2. Thanks Yes, fine. If the installer does not detect it automatically, try with es1371 or es1370 sound modules. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.4.0 #2 SMP Fri Jan 5 14:48:54 CET 2001 i686
Re: [expert] For all of you with One BIG anoying Pain in the A**
On 2001.01.02 Larry Marshall wrote: Mandrake. By version 7 of anything things like print services should be stable, period. This is especially true if the features have been stable through the previous few versions and no increased facility is being provided by the change. Well, don't agree with that. I don't remenber exactly what were the printer related problems, but I suppose that nobody sends a mail to a list to say 'hey, cups work'. There can be many people that probably are printing using cups and don't matter about this list. CUPS developers can test its package, beta testers can help, but one can not get all the cases tested in such a big system. I the masses don't burn CUPS, anybody will know if it works. Following your ideas, you will be still printing ASCII or PostScript Level 1 (if it is not too much bleeding-edge for you...) Kinda takes the wind out of the sails of "Linux is more stable than Windows" doesn't it? How do you answer these people when they say "But 7.2 is available in a store; can it really be that bad?" And Linux is really much more stable that windows. All the problems I see recently in this list come from that big thing called KDE, that tries to emulate windoze not only in the look'n'feel but also in the bugginess and size. So the problem is a buggy gui (well, the problem is that the gui is what you see). I have seen much less complaints on Gnome. I don't envy the job that companies like Mandrake have but they're making these decisions based on short term marketing, not in order to stabilize their products. Do you mean shipping Win 3.1 ?? Death is a maximum stability state (no entropy...). plunge. In short, as long as distro people are driven by marketing (and what else would drive them?) they will be adders/changers, not subtracters/stabilizers that's a fundamental problem in my view. I think that distro people have a very important role of 'choosers' of what soft are going to include. No distro can be equal to another. See Mdk and RH. Cousins that take very different ways. That is what I like of Mdk. RH7 could have offered a brand new buggy compiler, but XFree 4.01f was still a dream in RawHide when I donwloaded my new 4.0.2 in a working rpm from sunet. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3-aa3 #3 SMP Wed Dec 27 10:25:32 CET 2000 i686
Re: [expert] mkisofs
On 2000.12.28 Roger Sherman wrote: Prolly a bit OT, but I find I have to change most mp3 filenames...most of them are way too long to mess with if you do anything with them that involve a terminal... man mkisofs: - ISO cds allow up to 31 char names. But dos only reads 12 (8.3). So by default mkisofs cuts names to 8.3. - you can use -l option to allow the use of the 31 chars. But be warned that the CD perhaps is not readable in DOS. - you can use -J to store also Windows format names in the CD. That allows 64 chars. So names will look good (up to the 64 chars) under windows, good under Linux if the cd is mounted with Joliet support, and bad under dos. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3-aa3 #3 SMP Wed Dec 27 10:25:32 CET 2000 i686
Re: [expert] Disappearing Gnome
On 2000.12.28 Traci Collins wrote: Hi! I have a perplexing problem with a Mandrake 7.2 installation. Everything seems to go well, I have installed both Gnome and KDE but always boot graphically to Gnome, until I have a reason to reboot. Obviously I can go some time without needing to do that GRIN. But, when I do, I have a problem. Upon reboot I have lost Gnome support. See what /usr/sbin/chksession -l says. If it does not have Gnome, some thing is broken in your install. mdk rebuilds gui available envs each time it boots. Look at /etc/rc.d/init.d/mandrake_everytime It calls /usr/sbin/fndSession, which in turn calls /usr/sbin/chksession. All of them work based on some mdk specific files (eg, list of wm installed), that must be updated each time you insall a wm. So if you install an rpm from rhat or helix, perhaps it does not all the things needed to let the system knoe theres a new wm. (sorry for the typos and lack of caps, i'm typing one-hand with my 4 month daughter on the other...) -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3-aa3 #3 SMP Wed Dec 27 10:25:32 CET 2000 i686
Re: [expert] Dialup Trouble...
On 2000.12.26 Mark Weaver wrote: That is correct ergo the reason for using chmod 4755. that way nromal users can access the file owned by root. -- .. On Tuesday 26 December 2000 05:02, you wrote: I believe you must be *root* to ifup or ifdown. Pj Hey all, I've just started using the ifup and ifdown scripts instead of KPPP for dialup. Just one problem though...I get an "Users cannot control this device." error every time I try to run the scripts. I've done a chmod 755 on /dev/modem and /dev/ttyS1 as well as each of the scripts. Any ideas? You should not use that. That are to be called by higher level scripts. You should be using /etc/ppp/scripts/ppp-on and ppp-off (perhaps yours are named just pppon and pppoff), and they set up everything and call ifup-ppp. If you use that you just have to setup options in /etc/ppp/options. One other issue is who is allowed to dial. I think it can be done cleanly through the usermode tools, but I never get involved with them. There are two ways to do that: - Make the modem device world writable: chmod 666 /dev/ttyS1 (not 755, you don't have to EXECUTE(run) the device, just be able to write to it in addition to reading, to configure it and send init strings). - Make /sbin/pppd or /etc/ppp/scripts/ppp-on (off) suid (chmod +s .) I think the first way is less 'unsecure'. I now have a cable connection, so I don't remember if both things were needed. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3-aa1 #1 SMP Tue Dec 26 00:05:53 CET 2000 i686
Re: [expert] Uh-oh. Helixcode has not exactly improved my Foot menu...
On 2000.12.23 Simon Cousins wrote: In a fit of Xmas Cheer I have installed Helixcode's GNOME over a box running a schweet 7.2 system, via the Lynx go-gnome technique. The GNOME foot menu is now a mess. Helixcode has created a bunch of duplicate and mostly redundant menus. The mdk MenuDrake sees system menus only. The Helixcode menu util sees only it's menus. Ne'er the twain shall meet. So I can go in with vim and edit all menus by hand back to sanity, but if any lister had a better (ie, easier) solution, I'd be most appreciative. Helix Gnome 1.2 is fully included in Mdk 7.2. There can be a couple of packages not in mdk, but are neglectible. Even some of the packages are more up to date and bug-safe in mdk 7.2. Helix Gnome is well suited for old distros like mdk 7.1 or rh 6.2, that included gnome 1.0, but mdk 7.2 and rh 7.0 include full gnome 1.2 out of the box. Solution to your problem: do a rpm -a | grep helix and rpm -e all those packages or replace them with the correspondent mdk 7.2 package. Because this can be a mess, I suggest to you to boot in runlevel 3, deinstall any package named *gnome* or *helix* and reinstall the same from mdk CDs. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3 #1 SMP Fri Dec 22 02:38:17 CET 2000 i686
Re: [expert] Question about OpenSSL: RPM vs sources
On 2000.12.23 Michael O'Henly wrote: Hi... These install a few binaries, two libraries (libssl.so and libcrypto.so), and documentation. Dynamic libraries, loaded at runtime. OpenSSL sources (version 0.9.6) --- These install what the INSTALL doc refers to as "toolkits" (libssl.a and libcrypto.a), as well as a few binaries and documentation. Static ones, appended to program at compiled time. So your apache works because has been linked with the static libs, but kde can not find the dynamic ones it expects to load. Try configure'ing the build of OpenSSL with a flag like --enable-shared and --disable-static. See configure --help for the exact name of the option. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre3 #1 SMP Fri Dec 22 02:38:17 CET 2000 i686
Re: [expert] cups not working properly
On 2000.12.21 Praedor Tempus wrote: I am running KDE 2.1, cups-1.1.4-7.1mdk I thought I had cups setup properly for my Epson Color Stylus 600. I can print out test pages and I just succeeded in printing (sort of) an email from kmail - sort of because it all ran off the top of the page and didn't print all the message text - all it produces is the KDE graphic and the From and To email addresses. Message body text isn't produced. In spite of this very limited "success" printing, I find it completely impossible to print from konqueror. I tried repeatedly to printout a webpage I was reading but nothing ever comes of it. I checked the cups error_log file and it contains: E [20/Dec/2000:21:20:20 -0700] PID 16396 crashed on signal 11! E [20/Dec/2000:21:20:20 -0700] PID 16397 stopped with status 1! This is produced each time I try to print from konqueror. So, cups isn't working properly. How does one fix this? Every time I attempt to print from konqueror, konqueror freezes up for about 30 or 40 seconds before cups gives up. Don't be su sure it is cups fault (and after reading in the list all that about kde and konqueror, I would vote for konkeror). Check cups apart from any gui program. For example, you can get the example PostScript files that come with ghostscript (in my box they are at /usr/share/ghostscript/5.50/examples) and print them directly with 'lp'. So you check just cups, the way it can handle PS files. If this works, your problem is kde not talking same lang than cups. -- J.A. Magallon $ cd pub mailto:[EMAIL PROTECTED] $ more beer Linux werewolf 2.2.19-pre2 #3 SMP Mon Dec 18 12:13:19 CET 2000 i686
Re: [expert] cloning hard drives
On 2000.12.20 Rusty Carruth wrote: Bug Hunter [EMAIL PROTECTED] wrote: Can I use DD to clone the entire hard drive? I have used 'cp -a' to clone a full mdk install. The only thing you have to do is a boot diskette to boot cloned system the fist time and re-run lilo. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux werewolf 2.2.19-pre2 #2 SMP Sun Dec 17 00:51:15 CET 2000 i686
Re: [expert] Mandrake in production
On 2000.12.19 Scott Parks wrote: A guy who works for me tells me that Mandrake can not cut it when it comes to production web work and he favors, very strongly, Debian. Telling me that it is the strongest for production environments. I have been using Mandrake for several years and enjoy it as a personal development platform, connection sharing in my house, mp3 jukebox, etc. But, I have never used it for a web production environment. Anyone have some thoughts on this issue? The machines are dual P3's with 18 gig drives in arrays, 2 gigs ram each. Multiple machines behind Cisco Local Director. Mandrake installs fine on each box, Debian is a bit more bothersome to configure, but I can not simulate the load of web servers in the shop. I really doubt that one distro was better than other. For example, if you want a web server you must worry about what kernel to use (for good drivers for disks and net), and what version of apache or tux choose, etc. I still don't know why people 'labels' distros as good for workstation and bad for server (and the reverse). It all depends on the default install. Mandrake could distribute an installation with the 'server' mode as default, and be catalogued as a 'Server version of mdk'. Any distro contains now almost the same soft, and anyone can be torn down to the minimal running soft needed to do the work and do not loose any cpu cycle (say wipe out gnome, kde,...) Diffs: - How hard is to get the distro installed without the 'superfluous' soft to act as a server. - How hard is to upgrade the installed soft itself - How hard is to get the 'unusual' soft running (masquerading, firewalls...) - How hard is to get the box secure. Looking at all those, i think Mandrake is the right one. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux werewolf 2.2.19-pre2 #2 SMP Sun Dec 17 00:51:15 CET 2000 i686 Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] How does one get konqueror to work with java?
On 2000.12.19 Praedor Tempus wrote: I need javascript. The pages I am unable to use at my university are an ordering site for DNA sequencing - it uses javascript up the ying-yang. Not even the latest Mozilla can deal with it and Konqueror fails just as miserably. No error messages, nothing, just...nothing happens. The other site I am unable to use properly is an Outlook web interface to my local school email account. None of the interface "button" work and I am unable to view my email. It is also a javascript issue. Would this be something like the good all days when HTML was a standard until M$IE and Netscape worked their own extensions ? Perhaps those sites have somthing embedded (some extension) in their JS that is M$ specific, and k and moz follow strictly the standard. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux werewolf 2.2.19-pre2 #2 SMP Sun Dec 17 00:51:15 CET 2000 i686 Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Getting Mails from this list twice
On Sun, 10 Dec 2000 15:57:16 Jeff Malka wrote: We all are! And since the traffic on this list is voluminous, this makes it doubly voluminous. I wish someone would fix it. I am thinkin it has something to do with the headers in the mails. For example, this reply was done in balsa with 'reply to all', and it copied the addr '[EMAIL PROTECTED]' both in the To: and cc: fields. So perhaps headers are saying that the message was sent both to '[EMAIL PROTECTED]' and '[EMAIL PROTECTED]' in distinct places that should not repeat the address. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux werewolf 2.2.18-pre25-vm #4 SMP Fri Dec 8 01:59:48 CET 2000 i686 Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Alsa and ES1688
On Wed, 06 Dec 2000 13:58:33 Dom Savage wrote: Hi all, got a problem getting my sound card working as it's in a laptop and it's not plug and play. It works fine with OSSFree, but I was trying to get the alsa modules from mdk 7.2 to work. So far everything works fine, as I can play audio cd's and wavs and the like, but .mid file state: open /dev/sequencer: No such device. Check all the compatibility with OSS in modules.conf. I have: alias char-major-14 soundcore alias sound-service-0-0 snd-mixer-oss alias sound-service-0-3 snd-pcm-oss alias sound-service-0-12 snd-pcm-oss alias sound-service-0-1 snd-seq-oss alias sound-service-0-8 snd-seq-oss Perhaps the last two are what you are missing...but i have not tested the sequencer. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] XFree 4.0.1 not cooperating
On Tue, 05 Dec 2000 02:13:09 Ken Thompson wrote: SnipBad attempt. As in MDK 7.2, 3.3.6 and 4.0.1 can live together, the file named XF86Config is for 3.3.6, and XF86Config-4 for 4.0.1. So you should put your new file in place of -4Snip How in the world does a fella get 4.01 installed, or running after 3.3.6 was installed during set up. I installed the rpm, there is a XF86Config-4 file in /etc/X11 but still starts and runs 3.3.6. Get the Mandrakes CDs and find all the XFree4.0.1 rpms. Install and replace all the 3.3.6 packages. That is enough. The only rest of 3.3.6 is X11R6-contrib-3.3.6-10mdk. In 4.0.1 it seems to be easy to configure multiple monitors. Have not done yet. One idea: work with the given XF drivers until you get multi-monitor working, then try to use accelerated vendor drivers for TNT and Voodo. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Hack Kernel - which version is it?
On Tue, 05 Dec 2000 23:39:21 Larry Tobos wrote: TRY uname -a it should print the name of the system, It just prints the name that the packager put in the kernel Makefile. And it usually is consistent with the name of the package. For example, take mandrake. Ship the 2.2.17 kernel. It is named 2.2.17-1mdk. They add reiserfs. It becomes 2.2.17-2mdk. After a while they upgrade it up to standard 2.2.18-pre15. But it is named 2.2.17-3mdk. In the case they consider that all the patches introduced in pre15 are valid. Perhaps even they do a partial-pre15 (we trust this one, but not that other). That is a thing that makes me nervous. With distros kernels one never knows what has installed. So i prefer to get the standard kernel and patch. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Joysticks
On Tue, 05 Dec 2000 18:52:42 Balaji Ramani wrote: On Tuesday 05 December 2000 12:24, you wrote: stephen boulet wrote: Yes, but it shouldn't really be necessary should it. I did a quick check and I didn't find very many -joy modules of any kind. cu;-) Spence The problem is that the emu10k1 driver has a separate joystick driver. Most of the soundcard drivers have the joystick driver built in. That's the reason you need the emu10k1-joy module. One other problem is what is the exact relation between emu driver in kernel and the emu driver from Creative (which seems to be the base for the driver in kernel, but that one in kernel misses joystick support). ALSA is another problem. The alsa drivers for emu10k have not joystick support. I have seen a patch around there, but I have not been able to find it again. BTW, I have tried the Creative emu10k1-joy with the ALSA emu10k1, and works. So the joystick driver is completely independen of the sound driver. Perhaps it should be included with the joysticks drivers section in kernel. Is there any specific list to send a patch or ask about how to do it ? -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Alsa - and the ac97
On Mon, 04 Dec 2000 12:06:19 Stew Benedict wrote: You might try using alsactl: [...] You can then just add one line in rc.local - alsactl restore In 7.2 it is done automatically. See: Launch 'alsamixer' in a terminal. It is not very fancy, but works fine. Adjust your settings. Save them (you need to be root) in the system wide config /etc/asound.conf with 'amixer store'. And here comes the tricky part to guess if you dont know it: using chkconfig or ntsysv or tksysv, make sure that the 'sound' service is off, and the 'alsasound' service is on. The latter is responsible for restoring mixer settings at boot. Just: chkconfig --list chkconfig sound off chkconfig alsasound on chkconfig --list That works fine for me. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] XFree 4.0.1 not cooperating
On Mon, 04 Dec 2000 02:10:11 Tim Faehnle wrote: Okay, Does Mandrake 7.2 come with 4.0.1 or not? I'm confused by Mandrake install. (it's buggy and ill-designed, but that's a different discussion) It comes both with 3.3.6 and 4.0.1, that is the part that can be confusing you. install is finished and booted, I type "XFree86 -version" and I get 4.0.1. WTF? XFree86 always give 4.0+ because it is the name for the unified X server under 4.0. In other words, there is no XFree86 program in 3.3.6, you have a bunch of independent X servers: XF86_SVGA, XF86_S3, XF86_I128... Nothing tells me if that's what I'm supposed to see.). Okay, then I just overwrite the old /etc/X11/XF86Config with this new one (making a backup of course and removing the .new extension). Bad attempt. As in MDK 7.2, 3.3.6 and 4.0.1 can live together, the file named XF86Config is for 3.3.6, and XF86Config-4 for 4.0.1. So you should put your new file in place of -4. Then I type "startx" and it says that it can't even recognize the keyword "ServerLayout", which happens to be the first keyword in the new XF86Config, and it fails. What did I do wrong? StartX just tries to launch /etc/X11/X as server, that is supposed to point to the right one (is a symlink). In your case it seems to point to some server in the 3.3.6 series, that fails to read the new config tags. Change it to point to the 4 series (ln -sf /usr/X11R6/bin/XFree86 /etc/X11/X). And check manually the settings (driver selection and so on) in the -4 config file. I don't know how good is the configurator to guess your hardware. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Finding files based on date
On Sat, 02 Dec 2000 05:05:41 Bob Puff@NLE wrote: Thanks for the reply, but that doesn't work. At least in the form: find /etc/* -amin 8 Seems to return some files that have been created long ago. The actual files do seem to change with time. Example: [root@main postfix]# date Fri Dec 1 22:55:43 EST 2000 [root@main postfix]# find /etc/* -amin 10 /etc/fstab /etc/host.conf /etc/hosts /etc/postfix/virtual.db /etc/postfix/sender_access.db /etc/postfix/header_checks [root@main postfix]# dir ^^^ Please, remember you are on a Unix system 'ls' gives by default the modification time, and the flag -amin looks for files 'accesed' less than n minutes ago. That includes acess for reading and making no change. Perhaps you should try with the following flags: - ls: no flags give the modification time, -u flag gives the acces time including last time accesed for read - find: -amin or -atime, access (including reads). -mmin or -mtime, modification See man pages for more details. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] ALCATEL ADSL-modem and linux
On Sat, 02 Dec 2000 23:51:39 [EMAIL PROTECTED] wrote: I checked the log and it said unable to establish a connection ppp0 -- /dec/pts/0 ^^ Check your pppoe config. That should be /dev/pts/0, i suppose... ^ -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Finding files based on date
On Sat, 02 Dec 2000 00:34:45 Kelley Terry wrote: On Friday 01 December 2000 02:15 pm, you wrote: Question: how do I do a locate (whereis), based on date? What I want to do is display ALL files that have been created or modified SINCE a certain date.. like 2 days ago. man find. more tight, try find -ctime or find -atime. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] XFree 4.0.1 not cooperating
On Fri, 01 Dec 2000 23:29:28 Tim Faehnle wrote: I would like to get my Viper 550 AGP working with Nvidia drivers, but I need to update XFree86. So, I installed XFree86 4.0.1 with the RPM included with the Mandrake disk (I used the "mandrake update" program or whatever it's called. I think the file name was something like XFree-server-4.0.1.rpm.) Then, I exit Xwindows and use "xf86config" to set up my stuff. I'm positive that all the info is right, but when I then try "startx" it fails to init because it had an "unknown error" with the dbe module (couldn't find it, I think, but not sure). Is there anything special I need to do? insmod maybe? Am I looking at the correct configuration file ("XF86Config", right? or is it XF86Config-4?) Is there more I need to install? You should upgrade all the packages related to XFree (XFree86-*), but the -contrib package. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Installing Driver Fro Adaptec AHA-1542CP
On Thu, 30 Nov 2000 10:28:50 root wrote: Thanks for your reply. I tried this but unfortunately isapnp does not scan the address I am using or in fact any of the other adrresses that my card can use(I guess my card is pre-proper isapnp). Your reply was a I don't know if that makes a difference, but my in my 1520 the bios can be disabled with a group of jumpers (they set the bios address or no bios). Perhaps with no bios active the isa scanner can't find it. now I'm happy. One has to remember to turn the scanner on if one doesn't want kudzu popping up in the hardware detection phase of boot but then that another issue!! I really hate kudzu and linuxconf. Are like that smart car travel computers that continuously say "you have your back door opened" when you help a friend to move his furniture.. I know what hardware I have. Just the kernel module loader with /etc/modules.conf and device majors is fine. But I understand it is helpful for beginners. Thanks Again Not at all... -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre24-vm #2 SMP Wed Nov 29 02:56:21 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] URGENT:X-server produces a system breakdown.
On Sat, 25 Nov 2000 01:37:42 fabian wrote: Hello, I have a SiS 6326 video card in my system running LM 7.0. It was not very well configured, so I followed the advices posted in the newbie list. I added to the Section "Device" the "pci_burst_on" Option in the XF86Config. At starting the X manager the system broke down. When I rebooted the computer and after fixing the file system Linux started the X manager and the system died again. I wish I could let the system know at booting time not to start the X server in order to delete the added line in the XF86Config file. At the boot LILO: prompt, answer: somthing like LILO: Linux 3 The name 'Linux' can vary depending on the labels for your boot kernels. The important part is the 3, to boot in runlevel 3 (no X). -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] soundcard problem...conflicting irq?
On Sun, 26 Nov 2000 16:31:57 praedor wrote: I have an AudioPCI es1371 card and a PCI 3com 3C900B card presently sharing an irq. The soundcard has worked fine in the past, no cricket noises, no problems. I have now found that the same problem occurs under windoze, which made me think the card went bad. I picked up a new card, a soundblaster PCI 128 (drats!!) which isn't well supported under linux, it turns out. Windoze had the same problem with it as with the Bad attempt for a change. A SB128 is just an Ensoniq1371 or 1373, mine has a 1373 but the driver is the same, es1371. So it is just well supported as the AudioPCI... -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] soundcard problem...conflicting irq?
On Sat, 25 Nov 2000 22:53:30 Alan Shoemaker wrote: JAjust so you know, my Creative SB PCI128 uses an Ensoniq1370 chip which takes a different driver (es1370.o). -- Alan Mess of versions... -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] gdm
On Sun, 26 Nov 2000 01:48:21 alann wrote: But I got used to helix and to ME, gnome without it is lacking a lot of things in which I learned to do.. But what are you talking about ? 'Helix' is only the group of packages to run gnome, and you have them all in MDK 7.2. And you can even update them through cooker to some more recent versions than Helix offers. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
ld.so+glibc22 [was Re: [expert] MandrakeUpdate]
On Fri, 24 Nov 2000 18:30:15 Vincent Danen wrote: On Thu Nov 23, 2000 at 09:13:46AM -0800, Jean Meloche wrote: If anyone did manage to download and install the ld.so that was briefly on updates, I would recommend downgrading to the release version of ld.so (ld.so-1.9.11-4mdk) or if you are using cooker, to the latest version in cooker (-7mdk I believe). I am using an up to date cooker. I have glibc-2.2-7mdk and when upgrading to ld.so-1.9.11-7mdk, every ldconfig gives: ldconfig: warning: /lib/libdl.so.1 appears to be for multiple libc's ldconfig: warning: /lib/libdl.so.1.9.11 appears to be for multiple libc's There are two versions of libdl: -rwxr-xr-x1 root root 9140 Nov 22 15:01 libdl-2.2.so* lrwxrwxrwx1 root root 15 Nov 23 01:13 libdl.so.1 - libdl.so.1.9.11* -rwxr-xr-x1 root root 6472 Nov 23 04:14 libdl.so.1.9.11* lrwxrwxrwx1 root root 12 Nov 21 15:36 libdl.so.2 - libdl-2.2.so* I have moved libdl.so.1.9.11 and libdl.so.1 to /lib/old and everything works. Should't the %post in the ld.so rpm detect that there is a libdl.so from glibc22 and dont install its own ? Thanks -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Mandrake Auto-Install replication
On Wed, 22 Nov 2000 23:38:42 BillK wrote: Used to have this on an old 486: the "Li" is actually a message saying "I got this far through the boot". There are fixes but I have not the time to search for them. Try "man 5 lilo.conf" for how to set the verbose flag for progress messages. I think I found the original solution in the RedHat knowledgebase. It was a list of causes for failing at each stage of the boot. Usually that Li msg is due to misconfiguring the disk in lilo, say: - Install the distro (so lilo) with LBA mode on, and the loose the BIOS lba setting by a reset, battery exhausted or similar - lilo.conf has a flag 'lba' or 'compact' or 'linear' that is not supported by your disk+bios combo. In Mandrake its easy to install in non LBA mode, and the check the lba flag in the bootloader installation window, so does not work... -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre23-vm #3 SMP Wed Nov 22 22:33:53 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Hard disk problems
On Tue, 21 Nov 2000 17:16:46 Tejinder Singh wrote: 2) Duplicate the contents if the existing drive on the new one. The new one may have extra partitions if the partitions on it are exactly the same size as on the old disk. 3) Take out the old disk, or maybe even leave it in there as an extra disk and place the new one in as the main /dev/hda boot disk. 4) Power On the PC and the system comes up exactly as it used to before the noise started (with some extra disk space). Am I hoping for too much or is there a utility out there that can help me do this. I have just done this, manually. The only tricky step is partitioning the new disk. I have Mandrake 7.2, and I used DiskDrake to partition-format the drive. I suppose RH has any GUI tool similar to do that. In the worst case, you would have to use fdisk/mkfs. The thing is to get your drive partitioned and mounted in, say, /disk. You can have a distinct partition scheme than your original drive. Just format all your parts and mount them in their right point under /disk. Say /dev/hda is your current disk, and /dev/hdb will be the new: /dev/hdb1 --- /disk /dev/hdb5 --- /disk/usr /dev/hdb6 --- /disk/usr/local /dev/hdb7 --- /disk/home Copy all your dirs from your actual root to the future root, but WITH THE -a flag (so device files, links, dates and so on are preserved). All BUT /proc. Make that dir manually with mkdir /disk/proc: cp -a /bin /boot ... /var /disk Make a boot disk with your current kernel (mkbootkdisk). If you have all your mandatory stuff (disk drivers) in kernel, not modules, just dd if=/boot/vmlinuz of=/dev/fd0. Shut down, remove your old disk (or swap master and slave if they are ide, or change scsi ids to make the new disk seen before the old), boot with the floppy and run lilo on the new disk. The advantage of this method is that the partioning scheme does not have to be equal to the old. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre22-vm #7 SMP Sun Nov 19 03:29:20 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Upgrading glibc on MDK 7.2
On Tue, 21 Nov 2000 23:02:42 Jasenko Blazevic wrote: I did the same thing and upgraded to RehHat 7.0, it worked so well that i formatted the drive and installed Mandrake 7.2, don't do this, new glibc is so buggy, I couldn't compile and start any new application not to mention stability I am running previews of glibc for Mdk from 2.1.97 i think, and not a hang. Now I have 2.2-5mdk, and all works fine. The only problem was gcc2.96. Initially it was a bag of bugs. It didn't compile the simpler thinks in C++. But the actual rpm works fine. And the problem upgrading gcc+libstc++ is that you have to also update many programs that use lbstdc++. So everything will work in an upgrade to the latest cooker, but be prepared for a heavy session of rpm'ing. It's the cost of being update-adict (like me). -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Linux 2.2.18-pre22-vm #7 SMP Sun Nov 19 03:29:20 CET 2000 i686 unknown Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
RE: [expert] NVIDIA question
On Fri, 17 Nov 2000 18:45:52 RaEl wrote: NVIDIA has their own driver. Very easy to insstall. Just substitute XF86Config-4 for XF86Config in the instructions and all should work fine. I'm playing Quake3 and Unreal Tournament with no problems at all. As far getting a V5500 I ask why? A Geforce2 MX costs about a 100-120 bucks for a 32 meg card that pounds the voodoo into the dust in both windows and Linux. I used to be a Voodoo owner V1/V2 but Nvidia gets my cash because there cards are the best right now. Instead of using nvidia's card, try 3dfx voodoo 5500 card, it works great on my linux box running mandrake 7.1. Perhaps nVidia is not the best "open-source-friendly" company (although they are doing a great job with their drivers and the community, pitty that 'little' part they have to ship in binary...), but in hardware they are the best in normal budget market (don't think about Oxygens or TDZs of $1500). Beware this 3D acceleration support is experimental and is based upon previous version of nvidia drivers for RIVA128 and bug fixed by Utah GLX team. this version does not include any AGP acceleration, and really works almost correctly on RIVA128 card, but newer card may have problems with some games. Drivers from nVidia work fine. And installing is straightforward in Mandrake (it comes with XF401, Mesa split in Mesa- and Mesa-common). Over a sane install of Mandrake 7.2, you can get nVidia's drivers up on 5 minutes (15 if you have to read the docs) In my dept. we use Performer, and its a pleasure to see a GeForce2 DDR running at 100 fps on performer.town, at 1280x1024x16, and at 150 fps on a 2GTS (my home TNT2 only gets 30 fps at 32 bits...jealous...). -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Has anyone ever clustered Mandrake servers?
On Thu, 16 Nov 2000 23:11:47 James Little wrote: We have 4 identical poweredge Xeon 550 machines. They each have two network cards, and are on their own 10 port switch. The database we had them running has now moved to a single quad processor 733 machine. They're sitting there now with NT 4.0 advanced server doing nothing. Anybody know how to cluster them running Mandrake or RedHAt. I read everything I could find, but it's foggy at best. In rpmfind.net you have the kernel-mosix and mosix-utils packages, ready for building a mosix-type cluster. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] in which RPM can I find update-alternatives
On Thu, 16 Nov 2000 17:18:04 Jon wrote: update-alternatives). Now I'd like to know if anybody knows where I can find this binary or a RPM which includes this, or another way how I can get opensshd working again. werewolf:~ rpm -qf /usr/sbin/update-alternatives rpm-4.0-4mdk So I suppose it comes with a recent version of rpm, and is missing in your rpm in the 7.0. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] AMD motherboards
On Thu, 16 Nov 2000 23:46:29 Sridhar G wrote: Can anyone suggest me good Mb's for AMD 850 Socket A chip. Choose: http://www.via.com.tw/support/partnerskt133.htm -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] why does KDE change to user home for running apps ?
On Mon, 13 Nov 2000 05:13:44 Asheesh Laroia wrote: You could, alternatively: Boot up into runlevel 3 ("linux 3" at the lilo prompt) log in to Linux on a text console "xinit" At the XTerm you now see, cd to the dir you want and "startkde" Then minimize the XTerm. But if you launch anything (say another kterm) from a menu in KDE, the launcher does not know anything about which dir you started KDE from, justs starts a new terminal, that starts a new shell which is born in the home dir of the user. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Geforce2, NVidia drivers, and LM 7.2
On Mon, 13 Nov 2000 16:18:30 pgeorges wrote: Tom Berkley a écrit : ls /usr/X11R6/lib/modules/extensions/* /usr/X11R6/lib/modules/extensions/libGLcore.a* oops /usr/X11R6/lib/modules/extensions/libglx.so@ /usr/X11R6/lib/modules/extensions/libdbe.a* /usr/X11R6/lib/modules/extensions/libglx.so.1.0.5* /usr/X11R6/lib/modules/extensions/libdri.a* /usr/X11R6/lib/modules/extensions/libpex5.a* /usr/X11R6/lib/modules/extensions/libextmod.a* /usr/X11R6/lib/modules/extensions/librecord.a* /usr/X11R6/lib/modules/extensions/libglx.a* ^ oops again... /usr/X11R6/lib/modules/extensions/libxie.a* It seems as you have messed XFree (Mesa) GL-GLX and nVidia's one. Please, could you send the output of the followig commands ?: ls -l /usr/lib/libGL* /usr/X11R6/lib/libGL* ls -l /usr/X11R6/lib/modules/extensions -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Removing Extraneous E-mail Header Information
On Sat, 11 Nov 2000 06:44:04 Arcana wrote: Hello! I want a program that will save large numbers of E-mails into one large file for the purpose of archiving messages from a Mailing List in my inbox and putting them on a web site in a Plain-Text format. Try: http://www.hypermail.org/software/index.html -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Cable Modem Connection 7.2
On Fri, 10 Nov 2000 15:00:06 Wayne Stout wrote: I had a few problems with my cable modem setup as well, but mine all turned out to be caused by me. I found that when I had set up the eth0 interface under linuxconf, I had specified a different IRQ than what the isapnp file was setting up. As soon as I fixed that, (and made sure /etc/resolv.conf was correct) everything works perfectly. I had problems, but it was in 7.1. I reached 7.2 through periodic updates. My problems were related to the scripts for network start/stop in /etc/sysconfig/network. If you select BOOTPROTO=dhcp, no problem. But 'pump' seems much better than 'dhcpcd'. But if you select BOOTPROTO=pump, the start launches pump, the stop tried to kill dhcpd. So I finally rpm -e dhcp*, and let pump alone. Works fine. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Zip drives and supermount in MDK 7.2 does not work.
On Thu, 09 Nov 2000 20:26:42 Norman Carver wrote: Even without the icon, if you just cd into /mnt/zip in shell window, and then cd out or even close the shell window, pushing the eject button on the zip does nothing. I you can't spit the zip, try with 'lsof +D /mnt/zip' to see who has something opened in there. Perhaps kfm or gmc... -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] SMP
On Mon, 06 Nov 2000 17:38:13 Jeff Malka wrote: On mine too. Would not bootup that way till I changed lilo to boot to the up option. Maybe that is why it is not working correctly on my PC. Do you know how to force it not to load the smp kernel? Watch out for your motherboards. Some require the empty processor slot in a dual mobo to be "terminated" with a little circuit. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Kernel 2.4.0-test10
On Sun, 05 Nov 2000 02:32:38 Jason Straight wrote: I know this isn't Mandrake specific, at least I think not - but I tried to compile and run the 2.4.0test10 kernel (which I have compiled my own 2.4.0's from test3 without problems) and it get's to Uncompressing kernel ...ok booting linux kernel Then it just stops. Anyone know why it would stop here or somewhere I can go to get information on specifics like this? As a fast try, disable MTRR and recompile kernel. Look in the kernel mailing list aschives in http://www.tux.org/hypermail. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Cannot Mount / (root) via NFS
On Thu, 02 Nov 2000 17:06:17 Jeremy Schaeffer wrote: Hi! I cannot seen to find anyone that can answer why I cannot mount / (root) via NFS on Mandrake 7.0/7.1. I have three 7.0 Machines, one with a 2.2.16 kernel , one with a recompiled kernel with CONFIG_ROOT_NFS turned on, one with the stock kernel and a 7.1 with the stock kernel. None of them let me mount / , I get (logged in as root) mount -t nfs musha:/ /mnt/misc mount: musha:/ failed, reason given by server: Permission denied I suppose that server is refusing to export /. You need some other user space utils, to share roots between boxes. Even I don't know if it is possible. "root on nfs" means that you can have a diskless client kitty that mounts its root from musha:/export/kitty_root, and hope mounts musha:/export/hope_root. That needs net and nfs on the kernel. And is the easy setup. Suppose you want to mount musha:/export/root_for_clients both in hope and kitty, a shared root is not directly possible, there are files that are not read-only and both clients try to write in them (/var mainly). And imagine both the clients and the server at the same time. So the I have seen a special setup tool (i think it includes kernel and user space tools) to build a cluster of diskless nodes. See at http://www.beowulf.org or http://www.mosix.org. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Sprint Broadband internet
On Fri, 03 Nov 2000 04:11:06 Sridhar Govindarajulu wrote: I am planning to set up a hub/switch. In that case do I need another eth card. Currently I have a win client connected to my server. Perhaps you don't. If you don't want to make an internal network, and just want to connect both clients to the net. The manual in may cable-modem (Samsung) says that you can connect the hub directly to the cable-modem (with a crossover RJ45 cable), and the all the computers to the hub. They boot and request independent DHCP addresses. And a note: in Mandrake I have tested both 'pump' and 'dhcpd', and like very much the first. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] VGA parameters at boot
On Fri, 03 Nov 2000 07:37:53 Marco Fioretti wrote: Hello, I am very interested in having my PC booting right into a high resolution mode, because I have a fixed frequency monitor. Right now I have runlevel 5, but this means that I cannot work in text mode, and that if some error message appears during boot I never see it. I know about the Svgatext See /var/log/dmesg, and /var/log/boot.log mode package, but: 1) booting straight with a good VGA mode would give me readable text before than SVGATextmode could ever do, and would also be easier to set up, wouldn't it? I hardocde vga=6 in /etc/lilo.conf. It gives a fine 80x60 standard mode. 2) If this is the case, and assuming that my available mode are something like : 0 0F00 80x25 1 0F01 80x50 2 0F02 80x43 ... how can I figure out which of them is OK for my monitor (of I would try only text modes supported by standard SVGA. If you just want space to see messages and listings, that works fine. If you want to do graphics or many colors in console, try a graphic mode. But text modes will work without any fb support. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] VGA parameters at boot
On Fri, 03 Nov 2000 16:48:53 Alexander Skwar wrote: So sprach Marco Fioretti am Fri, Nov 03, 2000 at 12:41:16PM +0100: I have a fixed frequency monitor, so I have to be sure that the mode is right before doing that. What I was looking for is If I'm not totally wrong, the actual resolution (x Pixels by y Pixels) doesn't change at all, no matter what text resolution you choose. IIRC the text resolution just changes the size of a font while always staying at 640 Pixels x 200 Pixels (or something like this). So, if I'm correct, if one text resolution works, all the standards (80x25, 80x40, 80x50) will work. Tricky to find, but got it. The problem is that all that modes are text modes for a fixed standard VESA resolution and refresh rate, I suppose that 640x480x60HZ. If your monitor does not support that rate, don't work (I fighted time ago with a Sony GDM from a Sun attached to a PC..., only worked at 1280x1024x8bits, that gave just the right dot freqs...) Without frame-buffer support, you need to find a VESA mode that fits your mon. Attacched goes the modeDB.txt from XFree sources (hard to find without downloading the 30Mb of sources), which tells the WxHxFreq (line 648) of the standard modes. If the mode supported by your monitor is there, find which VESA mode from ftp://ftp.vesa.org/pub/VBE/vbe3.pdf (also attached) fits, and get the number. Then give this in lilo.conf as vga=mode code. Warning: don't remember if vga= eats mode numbers in hex or dec, prefixed or non with 0x. Check for the correct syntax. If the native (and only mode) for your monitor does not fit any of the VESA modes, you will need full fb support. Compile it in the kernel, and read linux/Documentation/fb/*. There says something about adding 0x200 to feed the mode to lilo. I have more detailed info such as which bit depth is each vesa mode, but only in paper from an oldie S3801. If i can scan it, i'll send you a PNG. Hope this helps you. -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] VGA parameters at boot
Sorry, forgot the files. Here goes modeDB.txt, vbe3.pdf is too big, grab it at ftp://ftp.vesa.prg/pub/VBE/vbe3.pdf -- Juan Antonio Magallon Lacarta # cd /pub mailto:[EMAIL PROTECTED] # more beer modeDB.txt.gz Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Postfix with Cable
Hi, everyone. A little question about postfix in mdk7.2. I have a modem-cable connection without a fixed IP, and want all my mail that has to go to outer world to be sent as coming from my address at my ISP "user@isp". I can send mail through my ISP smtp server, but From: name is that of the user who wrote the mail. I want the From field to be the ISP address, to get answers there, it the mail comes from root or from any user... I have looked throug the Postfix docs, but In can't see anything that suits this, i.e., rewrite From: addrees ONLY for non-local mails Any idea ? -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Stupid Question
On Wed, 01 Nov 2000 04:29:12 Chris Spencer wrote: On Tuesday 31 October 2000 19:57, you wrote: Why do you need /usr/src/linux in the $PATH? Chances are the files that Finding includes has nothing to do with PATH. are missing are in some devel package. Do a web search for the missing file and see what package it looks like it is in. You can always check rpmfind.net for the most up to date package and build that *.src.rpm or install the binary rpm. Now try to rebuild wine. I have the file in question...and the package. The file is part of the kernel source. Thats not the issue. The issue is why it is when I try to compile a SRPM it is not picking anything up from /usr/src/linux. Check if you have /usr/include/linux symlinked to /usr/src/linux/include/linux werewolf:~# ll /usr/include/linux lrwxrwxrwx1 root root 26 Jun 8 02:09 /usr/include/linux - ./src/linux/include/linux/ Same with /usr/include/asm. Or you can have both dirs "real" instead of symlinked, that depends on the dsitro. If some SRPM needs including kernel headers, they use linux/header.h; If you have the above link, gcc finds the include under default include paths (/usr/include), else you have to manually add -I/usr/src/linux/include. (NOTE: in the latter form, the last "/include" matters; I don't have your previous posts, but you seem to be refering always /usr/src/linux only). -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] c++ standard header missing
On Mon, 30 Oct 2000 12:02:21 [EMAIL PROTECTED] wrote: #I check with webmin and kpackage to confirm this and my box have this package is this enaugh? or is there any more package i'm missing? - gcc-c++-2.95.2-7mdk.i586.rpm installed - libstdc++-2.95.2-7mdk.i586.rpm installed - libstdc++-compat-2.95.2-7mdk.i586.rpm installed if anything I'm missing here? any url to download? (perhaps a good c++ header and working with 7.1 ;)) libstdc++-devel-2.95.2-7mdk.i586.rpm. I would tell you that it is on your CDs, but after some previous mails in the list perhaps it is missing. Check if you can find it on your media or grab it at http://rpmfind.net -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] c++ standard header missing
On Mon, 30 Oct 2000 18:48:59 Ronnie Whipp wrote: "J . A . Magallon" wrote: If this applies, use kpackage to uninstall ALL egcs packages. Then, from the 7.0 disk install the whole lot again. You have to do straight egcs first to avoid dependency problems. After this, MOVE the header files in /usr/include/g++-2 to /usr/include. I may have to do a bit more fiddling first, but at least this gets the old `Hello World' program to compile and work. It's better than nothing. It is worst than nothing. You are mixing headers from egcs and compiling with gcc-2.95, for example. EGCS libstdc++ is version 2, and g++-2.95 libstdc++ is version 3. If you have correctly installed both egcs-c++ and gcc-c++-2.95, you should have both /usr/include/g++-2 for egcs-c++, and /usr/include/g++-3 for g++, and two binaries of libstdc++, one for each. -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] HELP!!!! how do you give a user rights to stay in his own home dir? How do you turn off services to just one user?
On Fri, 27 Oct 2000 20:14:12 Larry Tobos wrote: How do you give a user rights to make them stay in his own home dir? How do you turn off services to just one user? man bash, look at option -r and info under section RESTRICTED SHELL. Make that user login shell bash -r. -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] here is the $170 question ^__^
On Fri, 27 Oct 2000 14:38:43 andy b wrote: ok... I bought 127 megs of ram to add to my 64 megs my bios recognises 192 megs... but Linux only recognises 64 (even if I only put in the 128 meg module) it would be nice to triply my ram... any ideas? Try booting with lilo: linux mem=192M If it works, add it permanent in lilo.conf: append="mem=192M" -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] Linux Mandrake 7.1 and libstdc++
On Thu, 26 Oct 2000 18:20:49 Larry Marshall wrote: My question is: Is this a (stupid/simple) mistake from Mandrake, or it is done in purpose? For whatever reason, v7.1 doesn't come with the ANSI library stuff. You've found the library. Now you need the egsc++ tar that holds the include files. BTW, all that stuff's included in v7.0 and v7.2. My two-cd set of 7.1, downloaded and home-burned, contains: CD1 (MDK 7.1 Install): libstdc++-2.95.2-7mdk.i586.rpm binarys to be capable of execute things compiled with C++. The only thing you need if you're not a developer. CD2 (MDK 7.1 Extensions): libstdc++-devel-2.95.2-7mdk.i586.rpm Headers and generic links (libstdc++.so - libstdc++.so.X.Y.Z) to develop soft: rpm -qilp libstdc++-devel-2.95.2-7mdk.i586.rpm ... /usr/include/g++-3/iostream.h ... Both usable with the gcc-g++-2.95.2-7.i586.rpm that is in CD1. Which is the mistake ? -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] LM 7.2 and beyond (part 2)
On Fri, 20 Oct 2000 23:39:23 Austin L. Denyer wrote: I hate to compare Linux to Windoze (in general Linux rocks, Windoze quivers) but the 'add/remove programs' bit will automagically tell you if you need additional stuff in order to run an application (albeit only stuff on their CD) and will automagically install it for you at the same time. The same goes for the Windoze Update facility that upgrades/adds applications over the Internet. If Billy Gates can do it, I'm sure the geniuses (genii?) behind the various Linux distros can do it too... I heavily disagree on that. One of the nice features of linux is that one can almost know in what spents each byte on the disk. Look at one other thread in this list, one about /var/log full of .x.gz files. People notices that. Suppose a server in which the admin logs remotely. It has just the basic X installed. (S)He decides that needs a GUI editor (to be used remotely). Looks for one, sees gEdit (or kedit), and ends up ('automagically') with the disk plenty of gnome (or kde) stuff (s)he didn't really neded (just try xedit, or something just plain gtk). And think in people outside USA. There internet connection (local calls) are free. In Spain, for example, you can't say 'ok, spent this night downloading StarOffice updates' @ not-sustained-50kbits (if you are lucky) and @ nearly 1$/hour. If linuxes start to auto-install thins very heavily, you will end with an ILoveYou.rpm wandering over the net (rpm runs suid). I'm just very happy with the actual way, use rpm or any graphical front end, and look at the dependecies and decide if you want all that extra stuff or look for another thing. If you need extra stuff, download and save it (for the bad luck of a system crash, you don't need to download it again). That is one of the things I don't like of MandrakeUpdate. It wipes the new rpms that just installed. My point of view it that if you can learn what a control panel is, where is "My programs" and where is the menu in Word to insert a table, you also can learn what is an rpm, what is a dependency and what is a kernel module. -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] WebCam device driver autoload
On Tue, 17 Oct 2000 19:52:51 Buchan Milne wrote: Ok conf.modules experts: First of all (don't want to be pedantic, just want to help...), if you use a recent modutils package, change all your stuff from conf.modules to modules.conf and delete the former. How do I auto load the driver for the ov511 chip (common in webcams) for my Creative Webcam 3. Ath present I load is in rc.local with modprobe ov511 See kernel docs: werewolf:/usr/src/linux/Documentation# rgrep -rl ov511 . /Configure.help /usb/ov511.txt Reading /usr/src/linux/Documentation/usb/ov511.txt, the appropiate major is number 81, and the device is /dev/video[X], so (USB gurus, correct me...) just with: alias char-major-81 ov511 that wil load ov511 when any app (xawtv?) access /dev/video. I don't know how to deal with various video devices, you will need to configure the driver load by major and minor or by device /dev/videoX... Depending on module dependencies, you will need to load at boot other modules. Following ov511.txt, more info at http://alpha.dyndns.org/ov511. -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
Re: [expert] 7.2 RC1 and Nvidia driver installation
On Sun, 15 Oct 2000 21:53:36 Darin wrote: Hi, I'm trying to get Unreal Tournament running under LM 7.2 RC1 but I'm having a few problems. Does anyone have any useable instructions on installing nVidia drivers? The instructions on the nVidia site are cryptic at best. All I can get now is my monitor going into oversync when starting UT. I think that instructions in http://www.nvidia.com/Products/OpenLinuxDwn.nsf/xfree86_40bld095FAQ are very clear, but... First, I would recommend getting the tarballs instead of the rpms, the build is so dependent on your kernel options (agpgart ando so on...) Note: I got this drivers working -and fast- on 2.2 kernels on an SMP box with a THT2Ultra, including the 2.2.18-pre series. No success on 2.4. I don't know if it is an SMP issue on 2.4, something related with TNT2 or what... Compiler makes no difference on 2.4, both egcs1.1.2 and gcc2.95.2 lock the mahiche hard when starting any GL app (2D for X works OK). So if you are thinking about 2.4 kernel, try first on 2.2. Passes: - Get XF401 working with "nv" driver. - Decide if you want to try AGP access with nVidia's propietary AGP or with kernel AGP GART. Configure your kernel and build it. AFTER that, build NVdriver kernel module. Its build depends on gart being selected on kernel or not. This is why you must build your own NVdriver. A driver built with gart active does not work on a kernel built without gart, and viceversa. It is not selectable at runtime. If you change your mind about agp, rebuild NVdriver. - Remove /usr/X11R6/lib/libMesaGL.*,/usr/X11R6/lib/libGL.*, /usr/X11R6/lib/modules/extensions/[libGLcore.a,glx.a] - Put nvidia's libGL.so* libGLcore.* at /usr/lib (i prefer that location better than /usr/X11R6/lib, so when you upgrade XFree dont have to copy again libGL's, just remove those from /usr/X11R6/lib). Run ldconfig to recreate libs links and cache - Put nvidia_drv.o - /usr/X11R6/lib/modules/drivers - Put libglx.so.1.0.5 - /usr/X11R6/lib/modules/extensions, and ln -s libglx.so.1.0.5 libglx.so. Make sure you DON'T have glx.a or libGLcore.a lying in /usr/X11R6/lib/modules/extensions. That is a common mess... - Change your driver in XF86Config-4 from "nv" to "nvidia". Also make sure you Load "glx". - If you cant get the resolution you had with "nv" XFree driver try adding Option "IgnoreEDID" "1" in your Device section. If you cant get this working, send a XFree.log attached... Good luck. -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.
[expert] Kernel 2.2.18 and GCC versions
Hi, everybody. I have a little problem when compiling new kernels. Kernel 2.2.18-pre15 compiles fine under gcc-2.95.2. It is just plain 2.2.17 with Alan's patch to 18-pre15. I downloaded the gcc-2.96 rpms from rufus, and the compilation process breaks at: werewolf:/usr/src/linux# make bzImage .. make[2]: Entering directory `/usr/src/linux-2.2.18-pre15/arch/i386/lib' cc -D__KERNEL__ -I/usr/src/linux/include -D__ASSEMBLY__ -D__SMP__ -traditional -c checksum.S -o checksum.o checksum.S:231: badly punctuated parameter list in #define checksum.S:237: badly punctuated parameter list in #define make[2]: *** [checksum.o] Echecksum.Srror 1 make[2]: Leaving directory `/usr/src/linux-2.2.18-pre15/arch/i386/lib' make[1]: *** [first_rule] Error 2 make[1]: Leaving directory `/usr/src/linux-2.2.18-pre15/arch/i386/lib' make: *** [_dir_arch/i386/lib] Error 2 The actual code in checksum.S that breaks 2.96 is: #define SRC(y...) \ : y ; \ .section __ex_table, "a" ; \ .long b, 6001f ; \ .previous #define DST(y...) \ : y ; \ .section __ex_table, "a" ; \ .long b, 6002f ; \ .previous I suppose the var-args macro syntax is less permissive in 2.96... Then, because this versions of gcc are "not officially supposed to work", got the egcs rpms (only egcs and egcs-cpp; BTW, kgcc didn't work, cpp was named 'cpp' instead of 'cpp0' in /usr/lib/gcc-lib/i586-mandrake-linux/egcs-2.91.66, and -traditional flag looked for tradcpp0, which I symlinked to cpp0). But then kgcc stops at: werewolf:/usr/src/linux# make bzImage .. make[1]: Entering directory `/usr/src/linux-2.2.18-pre15/arch/i386/kernel' /usr/bin/kgcc -D__KERNEL__ -I/usr/src/linux/include -D__ASSEMBLY__ -D__SMP__ -traditional -c trampoline.S -o trampoline.o /tmp/ccy1Wjci.s: Assembler messages: /tmp/ccy1Wjci.s:884: Error: can't handle non absolute segment in `ljmp' make[1]: *** [trampoline.o] Error 1 make[1]: Leaving directory `/usr/src/linux-2.2.18-pre15/arch/i386/kernel' make: *** [_dir_arch/i386/kernel] Error 2 TIA, and sorry for the 'long long' message... (I will post this also in lkml...) -- Juan Antonio Magallon Lacarta mailto:[EMAIL PROTECTED] Keep in touch with http://mandrakeforum.com: Subscribe the "[EMAIL PROTECTED]" mailing list.