Re: [Cooker] cedilla (was: urpmi beautiful)
On most setings: meta-g is ç. Just do cat, then try all meta-letter combinaisons, and note the few you use in every day life. -- Xavier
Re: [Cooker] no more key use for ssh ?
Yes, if users are aware of changes :-) Check the changelog: (vdanen@logan: ~)$ rpm -q openssh --changelog [...] And use ssh -v to understand what's wrong. To go back to protocol 1, in ~/.ssh/config: Host * (or whatever you need) Protocol 1,2
Re: [Cooker] Why gcc 2.96 and not 3.0?
The reputation of being binary incompatible is mostly based on rumors: it affects only dynamicly linked C++ code; and this same incompatibility exists between egcs-1.1.2 and gcc-2.95, and will exist between gcc-2.95 and gcc-3.0. What about gcc-2.96 - 3.0 ? Did these $^@$%$!$ again change the mangling so that there is once again this incompatibility ? I do not like the way you say mostly based on rumors. I don't know what percentage of the people who use gcc use it for C++, but nowadays, a lot of people do C++, and you nearly always use dynamic libraries. So it is not such a side-effect as gcc developers tend to say. -- Xavier
Re: [Cooker] Why gcc 2.96 and not 3.0?
On Fri, Jul 20, 2001 at 09:19:10PM +1000, Geoffrey Lee wrote: egcs is not compatible with gcc 2.95.x. gcc 2.95.x is not compile with gcc 2.96. gcc 2.96 is not compatible with gcc 3.0. End of story. We have distributed egcs and gcc 2.95 before, we had incompatibility, why are you worrying so much? I'm not worrying. It's just that for LM 8.1, it would be nice to have gcc 3.0... And to have the c++ compatibilities libs (there was a link to a missing library in the 8.0 when it was first distributed. Don't know if it was fixed, we're using 7.2 for our work, and I didn't try with the 8.0 since that time) with gcc 2.95.x, so that you could upgrade your system without needing to recompile everything. Cheers, -- Xavier
[OT] Re: [Cooker] Reply-To field
http://www.unicom.com/pw/reply-to-harmful.html This does not makes sense **at all**. List are suposed to be public bulletin boards, not private chat rooms! There is one point which is correct is the above web page, it's: FROM http://www.unicom.com/pw/reply-to-harmful.html: Can't Find My Way Back Home It may be impossible to reply to the author of a message once the Reply-To header is munged. The Reply-To header was not invented on a whim. It is there for the sender of a mail message to use. If you stomp on this header, you can lose important information. There are good reasons why the sender might insert a Reply-To header. The sender might not be the original author of the message (the name that appears in the From header). If responses should return to the sender and not the original author, then the sender will insert a Reply-To header. Or, maybe the sender added a Reply-To because he or she cannot receive email at the account from which the message was sent. There are many good reasons to place a Reply-To header into a mailing list message. If the Reply-To is munged by the mailing list, the value provided by the original sender is lost. Reply-To munging can make it impossible to reach the sender of a message. Note that there is no answer to this precise point in http://www.metasystema.org/essays/reply-to-useful.mhtml But the main point is that mailing-lists are poorly handled by the low-level mail headers, and specific headers should be used. I have to change my From: header to write to some mailing lists because they only accept this From: for the posting. I also change my From: for normal messages because people don't use the Reply-To for addressbooks usually but the From. Therefore, people get [EMAIL PROTECTED] which is explicitely referencing a specific computer, whereas I'd like to have [EMAIL PROTECTED], which I can redirect to any specific computer. Basically, I'm on my computer (lpnp69), and my official email is [EMAIL PROTECTED] I'd expect my configuration to be: From: [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] Tomorrow, for some obscure reason (let's say I'm changing lab), my computer becomes cdfas1, and I have no more account on lpnp69. Then two problems arise: - I cannot send mails to the mailing list, I have to resubscribe - people browsing the archives and wanting to send me a mail (I do that quite often, browsing an archive, saw that somebody had the same problem, and mail him to know if he has found something if the answer is not on the list) cannot do it, as my official email [EMAIL PROTECTED] has been mungled It should be the work of mailers to provide functions to handle mailing-lists. Of course, as only a few do, Reply-To mungling is after all a good idea. But in a perfect world, it shouldn't (perfect world=world where every body would use good mailers, such has mutt which have specific code to handle mailing lists). End of the rant for today... -- Xavier
[Cooker] sorting of new ls ?
Hi, with the new ls from fileutils-4.1-1mdk, ls sorts by the first alphanumerical character, *but ignoring* a leading '.'. ls -a is now a mix of .something files and normal files. Is this the expected behaviour of ls ? lpnp69 ~ ls -A a aa .AbiSuite a.c accept.ps .acrorc a.fig.bak a.gif aires [...] It also ignores lower/upper case... -- Xavier
[Cooker] new galeon
Hi, I'm trying to update from galeon-0.10.6-1mdk to galeon-0.11.0-1mdk: root@lpnp69 # rpm -Uvh /mnt/nfs/Mandrake/RPMS/galeon-0.11.0-1mdk.i586.rpm error: failed dependencies: libpanel_applet.so.0 is needed by galeon-0.11.0-1mdk root@lpnp69 # urpmi galeon One of the following packages is needed: 1- nautilus-1.0.3-3mdk 2- mc-4.5.54-2mdk What is your choice? (1-2) 2 To satisfy dependencies, the following packages are going to be installed (26 MB): mc-4.5.54-2mdk libscrollkeeper0-0.2-1mdk sawfish-0.38-3mdk gdk-pixbuf-xlib-0.11.0-1mdk rep-gtk-gnome-0.15-2mdk scrollkeeper-0.2-1mdk control-center-1.4.0.1-2mdk gnome-core-1.4.0.4-2mdk mandrake_desk-8.0-11mdk gdk-pixbuf-gnomecanvas-0.11.0-1mdk galeon-0.11.0-1mdk Is it ok? (Y/n) n root@lpnp69 # urpmf libpanel_applet.so.0 gnome-core:/usr/lib/libpanel_applet.so.0 gnome-core:/usr/lib/libpanel_applet.so.0.0.1 I'm quite surprised to need a 26Mb upgrade with mc and gnome just to have an update of a web browser. Shouldn't this libpanel_applet be in gnome-libs if it is needed by other packages ? Cheers, -- Xavier
Re: [Cooker] excellent! no kdebase-nsplugin installed!
On Sun, Apr 29, 2001 at 09:51:03PM +0200, Alexander Skwar wrote: So sprach Jose am Sun, Apr 29, 2001 at 10:56:32AM +: No telnet by default is VERY GOOD, if your talking about newbies. Use ssh, I think it is installed by default. No, it's not!!! No telnet-server is good (maybe), but why exclude the client? Not every server is telnet enabled, you know. Of course, and telnet is very useful to test tons of thing. Want to test your mail server ? How do you ssh smtp ? Test a web server ? ssh on port 80 ? Try to configure a cisco ? For newbies, explaining why ssh is (sometimes) better than telnet is something to do. Hiding them telnet is VERY BAD. If they have a window with a shell, they MUST learn what is telnet. From the man of ssh: DESCRIPTION ssh (SSH client) is a program for logging into a remote machine and for executing commands on a remote machine. It is intended to replace rlogin and rsh, and provide secure encrypted communications between two untrust ed hosts over an insecure network. X11 connections and arbitrary TCP/IP ports can also be forwarded over the secure channel. ssh is much better than rsh and rlogin. In no way it is a replacement for telnet.
Re: [Cooker] Redhat-7.1 vs Mandrake 8.0
I know where they are, and I installed it in a second moment... Maybe for any unknown reason I deleted bind-utils during package selection?! I don't think so. As far as I remember my last installation of a 8.0, a lot of usefull packages were not selected by clicking on all the big icons for installation, and there was no all packages option. Selecting all the big icons endend with installing something like 70% of the packages (I guess), and traceroute and nc were missing (I didn't try ping or nslookup). Currently, all my boxes that could go to 8.0 are back to 7.2 as there is no libstdc++-libc6.1-2.so.3 in 8.0 (there is a symlink in egcs-c++ package but it points to nothing). As usual, never install .0 distribs, wait for a while...
Re: [Cooker] Redhat-7.1 vs Mandrake 8.0
I know where they usually are, and they were not there. I search even with find, all over the disk... nothing! Anyway I'm quite happy for I'm reading that someone else has this same problem. Nothing deadly, but should be corrected in 8.1 or 8.01 if possible ;P Yes, doing a simple install selecting every box possible but not going into individual package selection, one ends up with a distribution in which a few things that should be here are missing: - elm/mutt/pine (the mail readers used by most of the Unix using people I know) - telnet - traceroute - ntp Then, there are also a few packages that could be nice to see: - minicom - dump - lsof - tcpdump - nc - finger - xfig Of course, it is still possible to install them afterwards, but when you select everything in the installation phase, there are a few things you expect to have (in particular things such as telnet!!!). Anyway, a Install all packages button would be nice. BTW, the install didn't keep my network configuration, I had to do it again. Maybe because I have no DNS or gateway? (private 192.168.1.x LAN). Cheers, -- Xavier
[Cooker] 8.0 and old version support
Ok, 8.0 is out, everything is wonderfull, we're on cooker, everyone has the latest kernel, latest everything... However, with 8.0 out, a decision has been taken to stop security updates on old versions (7.0 and before). I would like Mandrake people to reconsider strongly this decision. There are people with not-so-recent computers on which an "old" Linux Mandrake is running. I have 2 such computers, running 7.0. They have more than 60 users, a web server, and a lot of non-LM rpms things (scientific applications). Installing a new distribution (upgrading) is never so easy. And I only have 2 computers, what about people with a lot of them ? Please, please, don't push too much for new versions. I know it is not so easy to maintain old distributions, but it is just security bug fixes. You can have one PC with the full RPMS, SRPMS and running each of the old Mandrake versions. You cannot just say "it is provided without waranties". People trust you... And remember Mandrake 7.0 is only from Jan 2000! That's less than a year and a half ago!!! Cheers, -- Xavier
[Cooker] 3c509 isa-pnp and kernel 2-4
Hi all, my box has a 3Com 3c509B etherlink III ISA-PnP ethernet card. It works perfectly with all the 2.2 kernels I've tried (2.2.19-5mdk used currently). However, when trying to boot 2.4 kernels, the network doesn't work and after a few seconds I get tons of dmesgs: eth0: Infinite loop in interrupt, status 2001. Trying to figure out the difference between both, I have: - kernel 22 eth0: 3c509 at 0x300 tag 1, 10baseT port, address 00 60 97 c5 c9 47, IRQ 11. 3c509.c:1.16 (2.2) 2/3/98 [EMAIL PROTECTED] eth0: Setting Rx mode to 1 addresses. - kernel 24 eth0: 3c5x9 at 0x220, 10baseT port, address 00 60 97 45 c9 47, IRQ 11. 3c509.c:1.18 12Mar2001 [EMAIL PROTECTED] http://www.scyld.com/network/3c509.html eth1: 3c5x9 at 0x300, 10baseT port, address 00 60 97 c5 c9 47, IRQ 11. 3c509.c:1.18 12Mar2001 [EMAIL PROTECTED] http://www.scyld.com/network/3c509.html eth0: Setting Rx mode to 1 addresses. 0x220 is the address of my sb board! I have in my /etc/modules.conf alias eth0 3c509, and if I put alias eth0 3c509 io=0x300 irq=11, it doesn't change anything. Any idea? Everything else seems to work quite well (cooker rsync of yesterday). Cheers, -- Xavier
Re: [Cooker] nedit and netcat
fixed in nc-1.10-13mdk Why has it been fixed in nc and not in nedit ? In the FAQ of nedit (http://www.nedit.org/faq/sect_bugs.shtml) they recommend renaming nc to ncl to avoid the conflict. It is named nc on all the other unices I've used, and so usefull. Please move back netcat to nc, and nc to ncl in nedit package. Cheers, -- Xavier
[Cooker] rpm -U and sysV init
Hi all, most of the time, when I upgrade a package (tipically a security upgrade), if it is a server it goes automatically into /etc/rc.d/init.d/* to be started at boot time. If I do an upgrade (meaning I already have installed the package) and removed it from the boot time startup, why does it put it back again ? Sometimes, you need some package containing a server without wanting the server to be launched automatically on startup... So you install the package, and remove it from sysV init sequence. Now, whenever I run a MandrakeUpdate, I have to do a ntsysv just after. Can this anoying behaviour be corrected ? Cheers, -- Xavier
Re: Fwd: [Cooker] File named 3
if [ "$SECURE_TMP" = "yes" -o "$SECURE_TMP" = "1" -o "$SECURE_TMP" = "YES" -o "$SECURE_TMP" 3 ];then and I changed it to: if [ "$SECURE_TMP" = "yes" -o "$SECURE_TMP" = "1" -o "$SECURE_TMP" = "YES" ];then not sure if this is correct, but now I don't get the "3" file showing up any more. You should use: if [ "$SECURE_TMP" = "yes" -o "$SECURE_TMP" = "1" -o "$SECURE_TMP" = "YES" -o "$SECURE_TMP" -gt 3 ];then -- man test
[Cooker] SSH update ?
Given http://195.42.162.180/advisories/OW-003-ssh-traffic-analysis.txt is there an openssh-2.5.2 for LM 7.2 ? -- Xavier
Re: [Cooker] 2 gig file size
On Thu, Mar 15, 2001 at 06:26:21AM +, Ray wrote: Is the 2 gig file size limit gone in the new 8.0 beta?? Another good reason to try XFS.
Re: [Cooker] SSH and Drake Tools
Hmm, justed tested a bit more, and it seems that this error happens with all X apps?!? echo $XAUTHORITY ? If it says $HOME/.Xauthority, and if you use openssh, then just correct your shell startup script not to replace the $XAUTHORITY when it is set (it should be /tmp/ssh-/...). I had to correct all the Mandrake 7.2 I installed because /etc/skel/bashrc has the bug of [ -n $DISPLAY ]. The bug is corrected in cooker, but if you have an old or a custom_and_buggy .shellrc, you'll have this bug. -- Xavier
Re: [Cooker] Updating Packages in 8.0
If you want to update to latest cooker, you can use 'urpmi --auto-select'. Could someone update the man-page of urpmi to reflect what is obtained with urpmi --help? This option is very nice and should be documented! -- Xavier
Re: [Cooker] gcc-2.96
For you who haven't read about the use of gcc 2.96 you can read the this by Linus Torvalds. http://boudicca.tux.org/hypermail/linux-kernel/2000week51/0868.html /Snaggen, will not be using Mandrake in the future I agree! I won't using Mandrake 8.0. Ok, we all love Linux (or have to cope with it), but please, don't always believe in Linus as if he was god. You pointed out an interesting message in a very interesting thread. Read ALL the thread going from this message down and you'll see that gcc-2.96 WAS an increbible instable compiler the day RH used it. Currently, it's just the best g++. The only thing you could argue about is the fact that as always the API are not compatible with other gccs. But the problem here is the "g" in gcc that makes any binary compiled on a computer unusable on another one (I'm joking, of course, but the change in C++ API at each gcc is really a pain in the ass for those who use external libraries). gcc-2.96 is really usable, and even if Linus doesn't like it, read comments from Alan Cox for example. -- Xavier
Re: [Cooker] rpmdrake
nice point showing the cost -9 for compressing. But what about *uncompressing*? That's what is important here! Well, I thought -n in gzip changed nothing for decompression, just for compression. I would expect to gain a little in -9 just because the file is smaller and I have less to read, but no other gain in time. Therefore, from the numbers above, I would expect a gain of .6%. Some more tests (presented as an array not to have tons of shell lines): time to time to compress (%)decompress (%) size (%) gzip -1 6.28s (49) 1.75s (106) 8645419 (108) gzip (-6) 12.90s (100)1.65s (100) 8014728 (100) gzip -9 37.85s (293)1.59s (96) 7959056 (99) (bunzip2) 73.59s (570)9.77s (592) 7281284 (91) Using gzip -9 instead of gzip ends with a 4% gain in time, 1% in space, and a loss of 193% in time for compression. This is a large loss for such a tool if you use it for Cooker, as you have to rebuild the database every day or so ;-) BTW, if you leave the file uncompressed, it takes you a few miliseconds to read it. It's just 2.6 times larger than gziped -9. Given the price of hard disks today, the huge size of the Mandrake distribution, and the fact that waiting in front of a computer is boring... Well... I would really think about it (an option marked "save space" that uses gzip -9, else use no compression at all). ;-) -- Xavier
Re: [Cooker] SSH and Drake Tools
Again, this issue also stands for QUITE some time now. When you ssh to a machine and try to start one of the Mandrake tools with X (DrakConf, draknet, ), you'll get an error message which is due to the ssh tunnel. Once upon a time there was a time when ssh'ing and starting Mandrake tools was possible - will this be fixed? Works perfectly well for me (I mean, I can use DrakConf. Then, whatever I select, it crashes, but the same way with and without tunneling). What error message do you get? Are you sure you don't have an old .bashrc containing a line like [ -n $DISPLAY ] ? -- Xavier
[Cooker] rpmdrake
Some remarks on rpmdrake, as asked a few mails ago. First, it's very long. For exemple, it uses gzip -9. What's the interest of such use ? [root@lpnp69 bertou]# cp /var/lib/urpmi/hdlist.Cooker.cz a.gz [root@lpnp69 bertou]# gunzip a.gz [root@lpnp69 bertou]# time gzip a 14.12user 2.30system 0:22.17elapsed 74%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (5246major+89minor)pagefaults 0swaps [root@lpnp69 bertou]# ls -la a.gz -rw-rw-r--1 root root 8014728 Feb 28 18:00 a.gz [root@lpnp69 bertou]# gunzip a.gz [root@lpnp69 bertou]# time gzip -9 a 40.20user 4.54system 0:56.05elapsed 79%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (5214major+89minor)pagefaults 0swaps [root@lpnp69 bertou]# ls -la a.gz -rw-rw-r--1 root root 7959056 Feb 28 18:00 a.gz For .6% of gain in space, you have to wait 150% more in front of your computer :( Then rpmdrake tells me I have to install other packages to go on, I say ok, and it fails because of dependencies of other packages (ex: want to install openssh-client, tells me I need also openssh-passphrase. I say ok, conflicts with openssh-server. I cancel, install client, passphrase, server, and then conflict with pam...). Couldn't rpmdrake, from the beginning, tell me I have to install all of them? Still, I don't understand how to do an "upgrade" of a package. I have an rsync version of cooker on an NFS mounted disk, and I can't use rpmdrake the way I used MandrakeUpdate. Some other remark : why does the window printing "Please wait, updating database for the ... media" need to be "sticky" in front of every desktop? I use fvwm2 with virtual desktop and I have this window in front of my screen while I'm doing something else, which is quite anoying (it was not the case with MandrakeUpdate). When you install something, it doesn't update in the list. I tried to unselect it and it segfaulted. Basically, I would say rpmdrake is an alpha software. Usable, but far from being as usefull as MandrakeUpdate was (but cooker version of MandrakeUpdate segfaults after a few seconds of loading new packages). -- Xavier
[Cooker] xv
I don't see xv in the contribs of cooker. From a previous installation, I see: # rpm -qi xv [...] This is the famous 'xv' by John Bradley. It is shareware, but we ship it with the permission of the authors. It is a graphics viewer for many file types, including gif, jpg, tiff, xwd, etc. It also have manipulation features such as cropping, expanding, etc. You don't have anymore the permission of the author? Else it would really be nice to have it in the contribs, isn't it? -- Xavier
Re: [Cooker] xv
No shareware in contribs, only free software. Is there a cont_nofree_evil_dontinstall_rib version of xv for cooker somewhere, or do I have to compile it myself? (I don't mind compiling it by myself but if possible, I prefer installing an rpm, easier to manage). -- Xavier
[Cooker] Cooker-i586 20010225 4:16
Hi, some remarks about current cooker : I'm concerned about the installation of vi. There are a number of points upon which I wouldn't agree : - put vim in /bin/vim and not /usr/bin/vim. It has "always" be in /usr/bin, and should really be in /usr/bin. - put links to /etc/alternatives in /bin (vi is the only package I installed to do so) I had a (lot of) crash(es) and had to use /bin/vi without /usr mounted. It failed, and of course I used /bin/vim-minimal, but in automated scripts or the like it would fail. I would really prefer vim-minimal package to install a /bin/vi, and vim-enhanced to install a /usr/bin/vim. Then, kernel-2.4.1-22mdk does not work with my eth0 3c509 ISA PNP card, as I said in a previous mail. I run kernel22-2.2.18-14mdk currently... Any idea ? Some things about mutt package. Could the compressed folder patch be added to mutt ? And also when I use mutt, the 128 charecters appear as '?'. Of course it's Ok in xterm, vi, or whatever application I use. It never happened to me before with other mutts (I always compiled them myself because of the compressed folder patch). Any idea ? At last, I tried to launch netscape to give the URL of the patch of mutt and the xterm in which I entered netscape disapeared. Surprised, I launched another xterm, another netscape and beep, my computer rebooted :( Included here is the log of what happened. (I have netscape-common-4.76-2mdk and netscape-communicator-4.76-2mdk) Appart from that, it works well ;-) -- Xavier a
Re: [Cooker] i like being insulted by rpmlint :-)
# look for Java location if [ -n $JAVA_HOME ] ; then JAVALOC=$JAVA_HOME else if [ -n $JAVA_PATH] ; then JAVALOC=$JAVA_PATH else if [ -n $JDK_HOME] ; then JAVALOC=$JDK_HOME else if [ -n $JDK_PATH] ; then JAVALOC=$JDK_PATH fi fi fi fi you should add a ' ' before ending ']'. Also, let me remind you that this script is WRONG: lpnp69 ~ /bin/sh sh-2.04$ cat a.sh if [ -n $TOTO ] ; then echo "toto" fi sh-2.04$ chmod +x a.sh sh-2.04$ ./a.sh toto sh-2.04$ cat b.sh if [ -n "$TOTO" ] ; then echo "toto" fi sh-2.04$ chmod +x b.sh sh-2.04$ ./b.sh sh-2.04$ export TOTO="" sh-2.04$ ./a.sh toto sh-2.04$ ./b.sh sh-2.04$ export TOTO=x sh-2.04$ ./a.sh toto sh-2.04$ ./b.sh toto [ -n $VAR ] is ALWAYS true, whatever the value of $VAR, as if it is not defined it will expand in [ -n ] which is true. So please, correct all these damn [ -n $VAR ] into [ -n "$VAR" ] which expands in [ -n "" ] which is false. -- Xavier
[Cooker] Failed install
rsync to proxad a few minutes ago. use of network.img installation by NFS. It didn't recognise my card, but I told it (3c509) and it was happy. Then standard pannel appears, and instead of opening the window with the list of languages to select, it just prints a blinking 3x3 pixels or so empty box. On tty4(?) tons of lines with: unable to create sub named "" at /usr/bin/perl-install/c.pm line 9 Everything else seems to work (I have the shell on tty2 and so on). On tty1, X had 2 error messages: error opening security policy file /usr/X11R6/lib/X11/xserver/SecurityPolicy (seems logical as /usr/X11R6/lib/X11/xserver/ doesn't exist) and GLib_WARNING **: getpwuid_r(): failed due to: No such user 0 (is it normal ???) Anyway, after a few tens of seconds of blinking, an error window appear on tty7 giving me the same error message than tty4, and it goes to next selection with the same bugs. Any idea ? Did someone break something ? -- Xavier
Re: [Cooker] Failed install
Please give the output of lspcidrake (or /proc/bus/pci/devices) Intel Corporation|440BX/ZX - 82443BX/ZX Host bridge (unknown ignore) Intel Corporation|440BX/ZX - 82443BX/ZX AGP bridge (unknown ignore) Intel Corporation|82371AB PIIX4 ISA (unknown unknown) Intel Corporation|82371AB PIIX4 IDE (STORAGE_IDE unknown) Intel Corporation|82371AB PIIX4 USB (SERIAL_USB usb-uhci) Intel Corporation|82371AB PIIX4 ACPI (unknown unknown) Symbios|53c875 (STORAGE_SCSI ncr53c8xx) SubVendor=0x1de1 SubDevice=0x3904 Matrox|MGA G200 AGP [Millennium] AGP (DISPLAY_VGA Card:Matrox Millennium G200 4MB) SubVendor=0x102b SubDevice=0xff00 maybe there's a more interesting error message before this: can you try "more /tmp/ddebug.log" on console 2 ? Yes. 1st point, please recognize "q" in "more" to quit !!! I had to list all the ^$ messages... Error message is cannot find /usr/lib/perl/site_perl/po.cz2 which is obviously missing :( Any way to build it easily ? (I need to install a cooker before tonight) -- Xavier
[Cooker] eth0 problem
Hi, I've finally managed to install cooker with some help from Pixel. (version 20010224 20:30) Everything is more or less ok, except network (too bad :() A few seconds after launching network, the console gets filled with eth0: Infinite loop in interrupt, status 2001. It lasts only when I service network stop. During the few seconds without the error message, the network doesn't work. My card is supposed to be an etherlink III ISA (3c509-combo) A try of Harddrake gave a error message about module isa_pnp not found. Could it be the problem ? Any help would be welcome. Please, answer also at [EMAIL PROTECTED] It's not that I'm not on cooker, but... I'm subscrib-ed with [EMAIL PROTECTED] as address, and lpnp69.in2p3.fr is the computer on which I try to install cooker, therefore without network ;-) So please, Cc to [EMAIL PROTECTED] !! -- Xavier
[Cooker] reiserfs and chattr
Hi, is there some kind of chattr for reiserfs ? I really need a journalized file system but with the possibility of doing chattr +i dans chattr +d. Any way to have the same behaviour ? -- Xavier
Re: [Cooker] sysctl()
Is there a new kernel planned to take into account the "advisory" http://www.securityfocus.com/archive/1/161764 ? (I mean a new kernel in updates for LM 7.2, as I guess cooker kernel is or will soon be up-to-date) Yes. It's being worked on right now. BTW, the standard LM 7.2 kernel does not have CONFIG_SCSI_MULTI_LUN selected, which is a big problem when you use an autoloader for DAT backup for exemple (a big problem = need to recompile the kernel. Not such a bad thing, but anyway). Could it be in the default kernel ?
[Cooker] sysctl()
Is there a new kernel planned to take into account the "advisory" http://www.securityfocus.com/archive/1/161764 ? (I mean a new kernel in updates for LM 7.2, as I guess cooker kernel is or will soon be up-to-date)
[Cooker] .so files and 7.2...
Hi... I have a problem with .so files on a 7.2 system. I have a big library of some hand made programs. I make a .a libfile, and it works perfectly. If I make a .so file, it seems to work, but a program using it receives a SEGV. If I make the .so file on a Mandrake 7.1, it works perfectly. nm gives a diff between the 2 .so files, and no diff between the 2 .a files. The diff is that in 7.1, there is another symbol, a: 3dc0 A _etext that is not in the .so file given by 7.2. Here are some more infos: [root@arthur Linux-i686]# uname -a Linux arthur.cdas.augernet 2.2.17-21mdksmp #1 SMP Thu Oct 5 13:03:01 CEST 2000 i686 unknown [root@arthur Linux-i686]# gcc -shared -v -o libmoClient.so MoMess.o MoBits.o MoPrint.o MoStore.o MoRcvMess.o Reading specs from /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/specs gcc version 2.95.3 19991030 (prerelease) /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/collect2 -m elf_i386 -shared -o libmoClient.so /usr/lib/crti.o /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/crtbeginS.o -L/usr/lib/gcc-lib/i586-mandrake-linux/2.95.3 -L/usr/i586-mandrake-linux/lib MoMess.o MoBits.o MoPrint.o MoStore.o MoRcvMess.o -lgcc -lc -lgcc /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/crtendS.o /usr/lib/crtn.o -- this .so file is not working The "bug" might be in my program, but some gdb stuff indicates it SEGV trying to address a space not allowed, that is allowed when statically linked. The disapparition of the "A _etext" in the .so file surprises me. If I use /usr/lib/gcc-lib/i586-mandrake-linux/2.95.2/crtendS.o from a 7.1 distribution instead of the /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/crtendS.o from the 7.2 distribution, it works... Let's conclude I am a bit puzzled... What is this _etext symbol missing from the .so library when I link with /usr/lib/gcc-lib/i586-mandrake-linux/2.95.3/crtendS.o ? Any comment is welcomed, -- Xavier
[Cooker] 7.2 remarks
(first, a remark on sympa configuration: I'm regularly uns-nospam-ubscribed from the cooker mailing-list. I guess it's because my computer is sometimes down for a few hours. If it's the case, could you modify the configuration of sympa to remove people from the ML only after a week or so? [the uns-nospam-ubscribed is for sympa, because this "£$^%$ robot sent me my mail back :(]) Everything was perfect for 7.2, only some details: - /etc/skel/.bashrc contains an explicit export XAUTHORITY=... which breaks the automatic X11 forwarding provided by openssh - /var/log/sudo.log is not created and root is spammed for that entry missing every day (should be a touch /var/log/sudo.log in sudo package) - there is no faxmaster in /etc/aliases, but /etc/cron.daily/hylafax pipes something to mail faxmaster - there is no /var/www/html/index.html, but only /var/www/html/index.shtml which makes htdig complain (ln -s ?) - for someone who asked for a low security on his computer, /etc/cron.daily/logcheck is maybe not necessary ? It should be launched probably only with security "medium" or better. Good work! A last thing, I've had some problems with 7.1b and reiserfs, files being badly recovered and filled with \0 on sudden reset of a computer (our system is living in dangerous conditions ;-) I hope it will be ok in 7.2 -- Xavier
Re: [Cooker] 7.2 remarks
[ -n $DISPLAY ] {
[ -f /etc/profile.d/color_ls.sh ] source /etc/profile.d/color_ls.sh
export XAUTHORITY=$HOME/.Xauthority
}
If you use ssh, $DISPLAY will be set. Then the rest won't be executed.
Or am I wrong?
Well, with -n the {} is supposed to be executed if the string is NOT
null.
In fact, the test is bad:
[root@brian MetaSu]# unset TOTO
[root@brian MetaSu]# [ -n $TOTO ] echo toto
toto
[root@brian MetaSu]# [ -n "$TOTO" ] echo toto
[root@brian MetaSu]#
So anyway, [ -n $DISPLAY ] is true whatever DISPLAY is defined or not
(all $VAR should be in "" for the case of empty strings).
I don't know what the correct thing to do is, maybe :
[ -n "$DISPLAY" ] {
[ -f /etc/profile.d/color_ls.sh ] source /etc/profile.d/color_ls.sh
[ -n "$XAUTHORITY" ] || export XAUTHORITY=$HOME/.Xauthority
}
--
Xavier
Re: [Cooker] [OT] frozen electrons!
Again! electrons do not die Really ? Just guess what they're doing at the LEP experiment in CERN ? An electron can die, killed by a positron, disapearing in a bunch of energy (gamma). BTW, you'll never be able to prove they cannot die. The current limit of the life time of an electron is 4.3 10^23 years (in my particle booklet from 96, don't have the 2000 under the hand). -- Really out of topic
[Cooker] Replication Floppy
Hi... I thought the replication floppy (RF) would be a nice thing to install a few identical computers, but it doesn't behave at all as I expected. I'm doing a network expert install, and I thought that booting on the RF would ask me for an IP, a machine name, and then go on, using the same partitioning scheme, installing the same packages, and so on... What it did was just to remember it was an NFS install, keep all the IP configuration (so I endend up with 2 PCs with the same IP in my LAN, still don't understand how the install could work), and then ask me to do again all the partitioning and the choice of packages. It would be very nice to have a RF keeping at least the list of packages you've chosen (even if I'm beginning to know well Mandrake's packages, it still takes me 30m to check/uncheck all these). Is it planned ? What's the use of the current RF ? Everything went fine for the install, except some gs rpms, but they have been updated an hour ago, I'll have to check again. BTW, kde-i18-British is installed even if you don't select kde. And this time my Elsa Gloria Synergy was detected, thanx Pixel ;) Good work :) -- Xavier
Re: [Cooker] new DrakConf...
Hi, I saw that DrakConf has been remodeled. It's to big, It doesnt fit my 14" inc screen (I know it's to small!!!) And when I resize the window verticaly it should apeare a scrollbar or it sould rearange the icons to use the unused part of the window... BTW, just one remark, please don't forget people with a 800x600 screen when you do the desktop. On a lot of laptops (and even on large screen PCs) it's still a standard configuration. Lots of Mandrake applications don't fit in a 800x600 desktop by default...
Re: [Cooker] 7.2 beta2 --/dev/dsp permissions
what are the correct permissions for /dev/dsp? I also have a problem. I have no [xkg]dm running on my machine. I log on a virtual console, and then issue a startx most of the time. Everything is good, and /dev/dsp becomes bertou:audio upon login. However, I sometimes log off from the console, staying in the X session (nohuped). Is there a way not to have the ownership of /dev/dsp to go back to root ? Another thing, the permissions are 640 when owned by root. If it were 660, I could just add me in the audio group and it would be Ok. Is there some reason for not having the permissions set to 660 ?
[Cooker] gcc-2.95.2-11mdk
lpnp69 ~ rpm -ql gcc | grep bin /usr/bin/cc /usr/bin/gcc-2.95.2 /usr/bin/gcov /usr/bin/i586-mandrake-linux-gcc /usr/bin/protoize /usr/bin/unprotoize lpnp69 ~ ls -la /usr/bin/cc lrwxrwxrwx1 root root3 Aug 30 12:33 /usr/bin/cc - gcc ln -s /usr/bin/gcc-2.95.2 /usr/bin/gcc seems to be missing
[Cooker] Latest cooker install
Here is a report of latest (rsync on proxad.net 1h ago) cooker: First, it worked ! No more SCSI aicxx bug which hanged the installation in the previous (2 days ago) version. This is a development/expert install, installing everything exept games. As I already said, my Elsa Gloria Synergy card is detected as a 3Dlabs Permedia2, and I end up with a low resolution install :( During installation, the progress bar at the top is never going to the end of the box (no problem, but not nice), and the evaluated time started from 16 minutes to end with 25 minutes, with a +1 every 3 minutes... It is systematic, is there any way to have a more accurate estimation ? Starting services: there are a lot of useless services (adsl, dhcp...) that should not be checked by default as I answered "no" to the corresponding configuration pannel (network). postfix and sendmail are installed and both are checked for being automatically started (only postfix should be enough). xinetd appears twice. Also, there are by default 4 web servers (apache,boa,roxen,zope) launched at startup. Is it really a good idea ??? :) X-configuration: after selecting my card, I tried the "test again" button and it didn't test anything, asking for starting X at boot and then going on, just like if I said it was OK (maybe I missed the button and click a little below the right entry, but I don't think so). Booting... Boot sequence is ok, but looks like any LM boot sequence. I thought I read some messages about a new boot sequence ? (I used lilo) rpm installed: all the locales as well as all the kde-i18n were installed. Is it a feature ??? DrakConf complains about "can't find any program" (they are here, I can launch them with command line). There is an error rotating logs (been here for a long time) as /var/log/sudo.log doesn't exist. An empty one could be created by sudo rpm... Next, I wanted to test exporting a ReiserFS on a Hardware RAID via NFS, as this posed problem with my old LM7.1... During a first transfer of a few 100Mb from an NFS client (LM 7.1) to the RAID, the NFS server machine (the cooker one) blocked during about one minute, screen fixed, no answer to ctrl-alt-F1, answering ping but not telnet, ssh or any service. After a minute, the NFS transfer was finished and everything went back... I transfered again the same thing a few times and sometimes the same bug occurs. Then, I rebooted the NFS server while the transfer was being done (reset)... The behaviour is much better than before, but not reproductible : instead of filling the unfinished files with "\0", reiserfs just deleted them on reboot once, and let them unfinished the second time. But it is still not NFS compliant, as the client thinks the files have been transfered and doesn't go on with unfinished files. Is there an easy way to have ReiserFS on RAID NFS compliant ? It is really important for me. Everything else seems to work. Good job :) -- Xavier
Re: [Cooker] Latest cooker install
the problem here is to know what "all" packages mean. Either we install everything and people complain, or we don't install everything and people complain :'-( I don't complain about everything installed. I just said that I don't see how my PC can run sendmail and postfix at the same time, as well as 4 different web servers. So install everything if it's what the user asked for, but choose one of them for the startup ;) (BTW, I didn't try but I guess that's what it does in a non-expert install, no ?)
Re: [Cooker] Bug inside Installation when using SCSI ! PLEASE READ !
tty2 is not available at that point. The system has not mounted the CD. So I can't make a "bug" I think the shell at tty2 is loaded in phase 2 . Well, I have done a NFS install, and when probing for the SCSI interface, it loads aic7xxx.o (lsmod prints it as unused). However, it detects my SCSI drive, but doesn't assign him a letter, complaining about ppa and talking (in tty4) about problems with IOmega ZIP drives (never had such a thing in my system). Then of course, when going to partitioning, it complains about no /dev/sda device, and nothing more can be done. Tried "bug" command, but it hanged by system :( BTW, my Elsa Gloria Synergy video card is autoprobed as a Texas Permedia II one, and I end up with an awful VGA screen for installation (it looks nicer even on low resolution laptops) :(
[Cooker] sound and harddrake segfault
Hi... I just installed the latest cooker (rsync.proxad.net of today) on a somewhat standard PC (PII 400, 2 years old, no fancy hardware, a SB16 Awe soundcard, or something like that). There are some bugs in the "importance" of various RPMs which make you install by default "netscape-russian, netscape-sparc-francais", all the locales, a lot of aspell and kde-i18n files that are marked "maybe" instead of i18n like some other translated files... BTW, I tried to remove tcsh and was surprised by the numerous dependancies... Last thing, the "kdelibs 1 compatibility" package could be in the kde repertory of packages instead of having an entry for him alone. Ok, now I have what seems to be a functionnal computer except the following : - gnuplot does not recognise "set terminal gif", which is anoying for rc5 stats programs :) (even after installation of libgd-1.3 from contribs) - I don't have sound support, with : /dev/dsp: No such device audio: No such device - sndconfig builds an /etc/conf.modules while it is /etc/modules.conf - harddrake segfaults and dumps a core while in main(). As it is stripped, no more info from gdb. I tried to launch harddrake with all the -without-xxx enabled, and it still segfaulted - I took my old config /etc/modules.conf and it works :) The config is really standard : alias sound-slot-0 sb options sound dmabuf=1 options opl3 io=0x388 alias midi awe_wave post-install awe_wave /bin/sfxload /etc/midi/GU11-ROM.SF2 options sb io=0x220 irq=7 dma=0 dma16=5 mpu_io=0x330 (given by sndconfig) BTW, is there a way to tell aumix not to open an X window ? (better than unset DISPLAY aumix). Everything else running fine :) Good work ! -- Xavier
Re: [Cooker] ext3
On Fri, Aug 11, 2000 at 07:24:42AM -0700, Chmouel Boudjnah wrote: ReiserFS doesn't seem to be as stable as it is claimed, What is your definition of not stable ? Well, I know people who were using ReiserFS for a big database and went back to ext2 and fscks because they had some problems (don't know more details). I heard Redhat doesn't plan to ship ReiserFS in 7.0 because it is not stable enough. Personnaly, I had some tests of ReiserFS on a hardware RAID system exported by NFS and had a lot of problems. One of them being having files full with \0 on a sudden death of the NFS server. I've been using ext3 for about a year and never had any problem... Actually i don't plan to integrate the ext3 in our kernel not is stable or not but ReiserFS is much more tested, work great on server and we didn't have any bug reports. We still keep our eyes on the Journalised FS like Ext3, JFS, XFS for the future. Ok, so here is a bug report: 250 Go hardware RAID exported by NFS, a client writing on the RAID through NFS, push reset on the NFS server. I did it twice and each time I had corrupted files full of \0. The machines are running Mandrake 7.1b. Ext3 would really be nice (even if it is not proposed to the end user, but to have it as a module) for its ability to go back to ext2. -- Xavier
Re: [Cooker] ext3
On Wed, Aug 16, 2000 at 12:46:42PM +0200, Thierry Vignaud wrote: what's more, ext3 only exists for 2.2.x for the moment (it would take some time to port it to 2.4.x). Linux kernels are 2.2.x for the moment... At least, jfs exists for both 2.2.x and 2.4.0-testX. and ext3 is half as slow as ext2 ... on 2.2.x, reiserfs is faster than ext2 (they've quite the same speed on 2.4.0-testX) Sometime you don't need speed, you just need safety.
Re: [Cooker] LaTeX french
I guess you should better use babel. Here is a rough base:
\documentclass[a4paper,12pt]{article}
\usepackage[latin1]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[francais]{babel}
\begin{document}
Ça marche? Bien sûr! La preuve: même les espaces sont bon(ne)s!
L'électromagnétisme est coupé étrangement, mais il ne peut pas
connaître toutes les théories.
\end{document}
Works with basic tetex from Mandrake.
Re: [Cooker] Re: [CHRPM] xv-3.10a-8mdk
No kidding. Free is important. We now have many other good stuff -- and I know what I'm talking about, as a long-run xv user.. Just one (general) remark. Maybe you don't care, but I think Mandrake might loose some of its users going this way. I use Mandrake for some computers I administrate as it is easy to install and has a lot of things we need (such as journalized fs, etc.) out-of-the-box. However, I'm becoming more and more woried (ok, I'm french, still some work to do with my english) about some choices in Mandrake. xv is a standard. Removing it would do more harm that anything else. Some default Mandrake configurations are also against standards. It took some time to understand why the scrollbar was enabled by default whereas xterm man was saying the opposite (/etc/X11/Xresources). Then I was quite surprised by the choice of "unset wrap" in default vimrc. Maybe it's better for newbies (but are there vim-using newbies ?), but it is something unusable for anybody not used to. These choices will make me think twice before installing a Mandrake. Of course, it's nice to have Mandrake becoming more than just a "better configured RedHat". But if it's becoming explicitly a "newbie configured Linux", I won't use it anymore. I guess something in-between is needed. My advice would be: "leave old things alone", meaning don't modify things that newbie won't use. Put kterm and eterm entries, but not generic Xterm entries which contradict the man. Don't put strange default behaviour for things such as vi that no newbie will ever use (a newbie using vi should know about line length). Don't remove standards that are still used on all other unices. Then, for all the extra-stuff, ok, let's go for blinking purple and pink Eterms! ;) And maybe it could be nice to have differents iso images for different users. I installed M7.1 with only the first CD (didn't want to burn the second one), and I was quite surprised libstdc++-devel was not on it. Maybe a good reflexion upon the 1000? packages and a more subtle organisation would be usefull. Is it necessary to install 3 different CD burners softwares ? And so on... Just my 0.02$
[Cooker] man update
The security update of the man package provides unbziped man pages, which conflict : bzcat: Can't open input file ./manpath.1.bz2: No such file or directory. man.1 should be bziped. -- Xavier
Re: [Cooker] proftpd more secure than wuftpd?
On Wed, Jul 05, 2000 at 11:43:36PM +0200, Frank Meurer wrote: I've follow the thread about wuftpd and proftpd. [...] On Mon, 3 Jul 2000, lamagra wrote: ___ http://lamagra.seKure.de: advisory #1 Advisory: misc. bugs Programname: proftpd Versions: 1.2.0 = pre10 Vendor: proftpd.net Severity: high (root shell) and low Contact: [EMAIL PROTECTED] [...] Read http://www.proftpd.org/proftpd-l-archive/00-07/msg00060.html for an answer about it. I hope the mandrake rpm will follow soon (will it be in the update directory for MandrakeUpdate to catch it ?). -- Xavier
[Cooker] pybliographer-1.0.0-1.i386.rpm
Now that there is pygnome-1.0.53-1mdk.i586.rpm and pygtk-0.6.6-1mdk.i586.rpm in cooker, it could be nice to put also pybliographer-1.0.0-1.i386.rpm (works very nicely, maybe only mv pybliographer-1.0.0-1.i386.rpm pybliographer-1.0.0-1mdk.i386.rpm is needed :)
[Cooker] initscripts-4.97-40mdk
/etc/sysconfig/network-scripts/ifdown-ppp lacks the line : [ -f "$CONFIG" ] || CONFIG=ifcfg-$1 and therefore doesn't work out-of-the-box.
Re: [Cooker] various comments
- my graphic card (Elsa Gloria Synergy) is not detected, and lspci gives me a "Texas Instrument TVP4020 Permedia 2". Anyway, I couldn't get any resolution to work, with XFree86 3 or 4. It works very well with 6.0, 6.1, and 7.0 Mandrake (XF86_3DLabs server, but the card is still detected as the TI thing). could you try and find what needs to be done to make it work? Ok, so my card is detected upon install as a Permedia 2 AGP (and I have a VGA install screen all the way, which is bad), and when I change it to elsa gloria synergy, "test configuration" just tells me "an error occured", and on one of the consoles I have a X11TransSocket... cannot connect error. I say quit, don't forget the changes, don't boot in graphic mode. After boot, ldconfig, vi /etc/inittab to go to level 5, shutdown -r now, and it works. (This with XFree 3.3.6, I didn't try again with 4.0)
[Cooker] various comments
Here is a list of a few problems I get with Mandrake 7.1b. I don't know exactly what beta it is, as I built the iso from a ftp site (ftp.ciril.fr). The last iso I tried was built yesterday, so I guess it is up-to-date (last package from May 28). - my graphic card (Elsa Gloria Synergy) is not detected, and lspci gives me a "Texas Instrument TVP4020 Permedia 2". Anyway, I couldn't get any resolution to work, with XFree86 3 or 4. It works very well with 6.0, 6.1, and 7.0 Mandrake (XF86_3DLabs server, but the card is still detected as the TI thing). - when trying XFree86 4.0 setup at install, it doesn't work because it tries to write some logs in /var/log/XFree86.0. I had to ln -s /mnt/var/log /var/log before going on. - the printer configuration doesn't allow for a direct postscript network printer, or did I miss something ? I have a nice printer plugged-in my switch, and it would be nice to configure it upon installation. - on a Dell Latitude CPt laptop (celeron), using XFree86 4.0 is much better than 3.3.6, but when it goes into suspend (apm), I lose the graphic display when I come back (tty black, nothing happening). Sometime everything is blocked. Needed a hard reboot. - BTW, the isomake.sh would be better with the folowing patch (allows to relaunch the thing in case there is a problem) : 101d100 mkdir -p $mdkLOCALROOT/cd1 105c104 mv $mdkLOCALROOT/mirror/* $mdkLOCALROOT/cd1 --- mv $mdkLOCALROOT/mirror $mdkLOCALROOT/cd1 and I had also to change: 73c73 lftp -c "open $mdkFTPSITE ; cd $mdkFTPDIR ; mirror" --- lftp -c "open $mdkFTPSITE\; cd $mdkFTPDIR\; mirror" maybe because of my lftp version ? # rpm -qf /usr/bin/lftp lftp-2.2.0-2mdk Let's stop here. Anyway, I'm very positive about Mandrake 7.1, and if XFree 4.0 works well, and ReiserFS too (seems to work currently), then it will be perfect. -- Xavier
