Pierre Baco

My favorite desktop environment ? Mmmh, a clean desk in front of a windows without curtains. No phone, no paper, except a good book. Maybe a glass of Bordeaux on my right (after 11:30AM), an ashtray on my left (anytime), or a cat (relax, not for the same usage...) And concerning my PC desktop, KDE/Aqua is the default choice. Gnome looks really nice (and smart), but I never had time to fix the (very) little problems I had with it, so I use "default' KDE (and the command line prompt when I temper with XFree...). But good old "bare" b&w X is fun too...

I guess they forgot to include hostname mgmt in MCC. Couldn't find it. No real speed improvement for me on my old Celeron sandbox, but all hardware recognized. Loading XFree is slow (although video card and screen correctly detected). Ah, and a really slow login time: takes 3 to 4 secs to get the prompt once you've entered the correct pwd. PAM problem ? During install, I've also found a couple of "bad" rpms on CD 3 and 4, despite a correct md5sum. As far as I remember, it was in OpenOffice and Mandrake french documentation. Apart from that, nice distro, and already 300 Mb of updates, found on the Netherlands mirror site (the french ones were not up to date).

Your comments on Bruce Perens new project ? http://www.userlinux.com/cgi-bin/wiki.pl?White_Paper

I've downloaded the kernel-source-2.6.0-1mdk.i586.rpm, in order to compile it on my 9.2 up-to-date machine. Well, make xconfig + bzImage + modules are okay, but I can't figure why "make modules_install" (i.e. depmod) keeps complaining about tons of unresolved symbols in major modules like ide-cd, ext3, ppp, isofs, sound etc. I'm not talking about third-party obscure modules there! Some errors can be resolved by compiling the option in the kernel, but that's not a solution. I want to understand WHY I get these errors. Actually I get same kind of errors when using the original 2.6 kernel source tarball from Linux Kernel archives, but not in the same modules. Did not happen to me since....8.1 kernel recompile. How can the kernel and modules always compile fine and depmod keeps on finding unresolved symbols ? Btw, when compiling the plain original MDK9.2 kernel with the original Mandrake .config file, (from /boot), I often get the same symptom. Looks like the default kernel/ default config can't be recompiled without errors. Any idea ? :blink:

It's the full 7 CDs stuff.

Same in France. Approx 8 Euros. Sounds reasonable for a Christmas gift...

I had a bad evening solving this f... security update problem. I managed to do it but it took me more time than expected. First I was really upset by yet-another-too-soon-released update, second by the lack of reaction in MandrakeClub and then I read your post here. It was a bit too much. I admit I'm not specially receptive on saturday mornings After 3/4 coffees and half a pack of cigarettes I feel better now. Next time let's sync our clocks: when you post something a bit sarcastic, let me know, so I won't read it before 11:00 AM ;) but keep on doing it, okay ? Just like I'll continue to answer on the spot! :D

Who's complaining "at every turn" and "ad nauseum" here ? Me ? I thought this forum was a place for Mandrake users to share their experiences with Mandrake Linux and get help from expert users. I did not realize I commited a crime when posting about a stupid problem in a security update. Oh, sorry, Master. But this did not stop me to switch, step by step, to linux three years ago. Are we supposed to show an eternal smile on our peaceful faces to keep the rest of the world unaware of the little bugs we encounter from time to time ? Mandrake has its flaws, like any other platform, and when there's a big one (like a security update breaking X), I'd rather inform the 'newbies' about it than keeeping my mouth shut. Who said I used MandrakeUpdate ? I don't. I prefer to use urpmi from command line so you can see exactly what's going on. And btw, urpmi did not complain when installing the update. When you receive a msg about a "security update" from your preferred Linux editor, you're supposed to trust him no? Unless you test all updates on a sandbox machine. Unfortunately, my other PC runs Debian... What do you think we did ? Format+reinstall from scratch ? If I was a real newbie, that's the kind of comment that would convince me to stay with Windows...Makes me feel very stupid and ignorant, trying to enter a world reserved to arrogant gurus.

What the f... is Mandrake doing with this 'security' update ? I installed it this afternoon. When restarting my PC tonight, I got an 'execve error 2' when calling startx! found that the /etc/X11/X symbolic link was dead because.... /usr/X11/bin/XFree86 was no longer there! Great! Went to MandrakeClub, where nobody seems to have noticed. Just like Chris Z, I had to do a "downgrade" of all XFree86* packages from CD1, using plain terminal urpmi. I suspect the xfree86-4.3-8.4.91mdk update deleted parts of XFree86-server and/or XFree86-server-common rpms. .

Got it! I've never been to those Office submenus, as I've simplified them to the bare minimal. Thanx, over.

Hi'vrybody. I'm looking for a small postit utility program. You know, those yellow sticky notes you can drop on your desktop (KDE, Gnome, I don't care). I had the name of one, so I wrote it down on a postit (a real on). But I lost it!

Looks like yours is already equipped for that!!

Well said! *nixes zealots should not use FUD. The "Linux über alles!" fanatics generate more trouble than the other side. Linux has flaws. And the best way to fight FUD is to tell the world about those flaws (no dirty secrets), how many (not so much), how fast there are fixed by really competent guys (very responsive community), and how quickly and easily everybody can then update their systems (no monthly giant service packs).

I think (and therefore I am) it's a very good idea. Wonder why such a thing does not exist yet. I admit it must be complex to manage, except for expert moderators like ours... First ideas: - a free virtual system like VMware. - a new device to replace keyboard and mouse (no speech recognition) - an horizontal, waterproof, ashproof, 3D multi layered display device emulating my desktop mess (i.e. MY way to sort things) -and a virus to kill ugly avatars

I approve 100%. I would add that the main problem for getting Linux on desktops is the lack of vertical apps. There a plenty of nice, good and rock solid horizontal apps, utilities and services. But computing is not an objective, just a tool to get things done. Unfortunately, major software houses designing vertical apps will not go GPL for a while. May take years to switch to another economic model based on services rather than royalties and charged licences. Meanwhile, I would not mind to buy, install, and promote proprietary vertical apps as long as there are running on top of GNU/Linux. We need accounting, payroll, inventory and all this sort of not-so-funny-and-sexy stuff on our desktops NOW. That's what I get from the professional users everyday. Sure, knowing what's under the hood is important, but you don't become a mechanic overnight, and more important, some of them don't want to become mechanics as they have other things to do, like running their business. They might be impressed and convinced by GNU/Linux quality and solidity (once they've tried it), but that's not enough to deploy it everywhere. I think we're at an important step, where free software is (almost) accepted as a very good way to build a serious computing infrastructure (network, servers, horizontal services), but it may stop here or stay as a "niche market" if developers do not switch to this desktop platform for their vertical apps. We've seen pretty good OSes disappearing just for this... This was my 1 euro contribution.

Let's say we write a common, public, official, stamped, signed and countersigned letter to the SCO case judge, saying something like: "Your honor, We, users of GNU/Linux, have choosen to use this OS for years because we believed the GNU GPL under which its code is released was protecting us against any copyright and patent claim from companies like SCO. As of today, we find ourselves in a peculiar situation where we could be sued for illegal practices like copying, modifying and distributing software code that does not belong to us and for which no rights to do such things have been granted. In other words, we might be sued for a "crime" we did not known we were commiting: under its terms, the GPL authorized us to do whatever we wanted. We are honest citizens, and if we are really breaking the law, we want to know how, where, and since when, in order to correct the situation asap. If SCO claim is funded, we need to know. Otherwise, how could we be condemned if there are no evidences of our crime ? Actually, we are the innocent victims of a very late SCO decision to claim property on parts of the GNU/Linux code. We are simple users and developers of software code, and it is not in our intention or interest to support one side or another in this trial. As an act of goodwill, and in order to save public money, we are ready to release to the court all the source code we use, and most important, all its history back to the original code writers. Would you be kind enough to ask SCO to do the same thing, so you can compare it to ours ? As regular users of the software involved in this case, and since we are the potential victims of the consequences of the trial between IBM and SCO, we are ready to help you on this matter. It is our interest to get ourselves out of this situation asap. This puts us in a perfect position to examine and verify the evidences. ...." Etc... I know it sounds pretty hyprocrit (knowing you free software zealots :P !). But look: in the french legal system, there's a thing called "partie civile". Anybody "indirectly" involved in a trial (like being a "second level" victim or accused) may ask the judge the have access to the evidences and eventually ask for a symbolic indemnification, or even sue one of the two trial parties. Imagine a group of 100 GNU/Linux users asking the judge to do so...We don't really care about IBM or SCO...Afterall, we have our own very private and egoist rights to defend in this case, no ? ;) Another way to have access to a trial files when you're not directly involved in it is asking the judge to be sued for the same crime, because you think you may have committed it without knowing it, and you want to prepare your defense for future eventual cases. Its very frequent here. I wonder it this is possible in USA. Might be a good way to win a free flight to America...But what about the return ticket if SCO wins ?

I've seen that Stallman has been aslo served with a subpoena in this case. IANAL, but I wonder what the SCO ones smoke to get such crazy ideas. Calling Torvalds and Stallman in court ? They're going to eat alive the SCO lawyers. One positive point anyway: For such a show, RMS may change his T-shirt for a clean one. For an haircut and a tie, no way. Unless SCO wants to patent the GNU GPL.

I can hardly see a suitable business model for Novell selling Linux shrink wrapped boxes. Novell (was and still) is a big company (revenue, customer base, # of employees...), mainly making money out of NetWare and network services. If they wish to increase their presence on the Linux SERVICE market, they did not need to buy Suse. Unless they want to release a Novell Linux, filled up with proprietary add-ons, plugins, drivers and services. Anyway, I've been working for Novell for 9 years. I remember that behind Novell was Ray Noorda (I wonder if he's still a big Novell shareholder). This guy created the company (first CEO). And since 1985, he has spend a fortune trying to launch desktop OS alternatives to Windows (including acquiring Digital Research, and many other graphical/windows systems to be included/melted with NetWare client software or handled by other Noorda's smaller companies). Maybe he's still fighting Msoft....With this kind of weapon (NetWare + Suse distro + Ximian) he may have a chance (but a proprietary one).

As they talk about "freedom of choice", "enpowering people" and "capability to move", I bet they're going to release a traveller universal power plug with a 5 meters extension cable... I can't wait Christmas to get one ! :wub:

Sorry for late answer: (02:15 AM) Just coming back from a seafood - cheese party !! :wink: Ok: I'm talking about fixing ACPI to run on a Compaq laptop 702EA under Mandrake 9.2 Powerpack. => Kernel is linux-2.4.22-10mdk. I know there's been another kernel update (Thursday or Wednesday), but I did not used it. What I've done to make it run on my machine: 1°) Create another /usr/src/ subdir for your own kernel version by copying the original one. Mine is linux-2.4.22-10yan. Link it to /usr/src/linux. Update Makefile to reflect the new name. 2°) Run (x)config to set/unset the following options: a) Processor Type and features: - Processor family => Athlon/Duron/K7 (for the Compaq 702EA) - CPU Frequence scaling => No - Toshiba & Dell laptop support => No - Symetric multi-processor support => No - Local APIC support on uniprocessors => No - Unsynced TSC Support => No B) General Setup - ISA, EISA and MCA support => No - Power management support => Yes - APM support => NO. Note: Setting it to NO will grey out an important option like RTC stores time in GMT, that should be left to Yes. c) ACPI Support - ACPI Support => Yes - Read DSDT from initrd => No - AC Adapter, Battery, Button, Fan, Processor, Thermal zone =>Module - ASUS, Toshiba and Debug statement => No - Relaxed AML checking => Yes Now, I've unset many other kernel options that I don't really need on my laptop. I can send you my complete .config file if you want. I've tried to set the ACPI options to Yes (built in) but I had strange err msgs in the syslog (like "ACPI: loading battery: FAILED and so on). Switching to modules solved the problem. Don't ask me why. 3) Run make dep, make bzImage, make modules, make modules_install and make install (will make the initrd.img file for you). 4) Edit /etc/lilo.conf to modify your new kernel options. Mine looks like (extract): image=/boot/vmlinuz-2.4.22-10yan label="9.2yan" root=/dev/hda2 read-only optional vga=791 append="devfs=mount acpi=force noapic resume=/dev/hda5 splash=silent" initrd=/boot/initrd-2.4.22-10yan.img Note: you may have to rename the corresponding vmlinuz and initrd files in /boot. (Can't remember if i did it or make install did it for me). 5) Run lilo to save-check-set changes. 6) Install the acpi-0.6-9mdk and acpid-1.0.2-4mdk packages. Use mcc to make sure acpid will be lauched at boot. 7) Reboot. It works (at least here !). The KDE latptop applet detects battery/power AC, battery level and sends alarms when needed. The fans and disks stop when on battery and not needed. Lid and button events are trapped (power button event calls the shutdown routine). I miss the thermal zone stuff, but i guess the Compaq 702EA is not built for this. Anyway, the laptop is much cooler than before. The acpi command displays things like: "Battery 1: discharging, 5%, 00:05:48 remaining." Not really accurate (pessimistic estimation, you'll need to add at least 15 minutes). Still working on power suspend (well, not too much so far). Well, that's all I can say right now. Hope this helps. Any Pepto Bismol around ? :(

Interesting but dense and hard to follow. Is he talking real english ? I mean for ordinary people like me ? Sometimes lisp looks more readable. There should be a law forcing lawyers and attorneys to use a max of 10 words per sentence, out of a 3000 words dictionary. You know, a kind of RISL (Reduced Instruction Set Lawyer)...

Well, it took me time. I had to recompile the kernel, remove APM, set ACPI for modules (not kernel built in support, modules ! don't ask me why...), remove CPU frequency scaling support, and set local APIC support on uniprocessors to NO. And I set the "load DSDT from initrd" to OFF. Then, I updated lilo for kernel options "acpi=force" (new 2.4.22 feature) and "noapic", and installed acpi and acpid rpms. Reboot and bingo ! /proc/acpi is fully populated now (although I've no fan and thermal zone support). Anyway, CPU and fan go to sleep gently when on battery only. And when I press the power button, "shutdown -h" is called. KDE battery applet works too (although the alarm is triggered a bit early). Still working on power suspend. The http://acpi.sourceforge.net site has very good explanations about incomplete/corrupted DSDT problems.

That's what I'm currently doing. I want to understand this usb stuff. Btw, I've found a modem cable: everything is fine. I can connect to my ISP using the old dialup line. Merci encore!!

What you sent me is the i686 version of what I had in 386 version (according to package names found in the 9.2 powerpack edition). 1°) it did not compile correctly (when compiler invoked by hsfconfig): syntax error in ../modules/osspec/include/oscompat.h,lines 274 and 320. (missing " character). I removed the extra " chars and it compiled ok. 2°) Compile error solved, but still exact same problem running hsfconfig: depmod: *** Unresolved symbols in /lib/modules/2.4.22-10yan/kernel/drivers/usb/usb-ohci.o depmod: register_resume_notifier depmod: suspend_task depmod: unregister_resume_notifier Looks like my usb_ohci module has a wierd problem. I'm gonna take a look at the src version of your hsflinmodem rpm to see why and when it needs usb support. Last minute: it works anyway ! :shock: hsf** are modules loaded. KPPP sees the modem (/dev/ttySHSF0) and I can send ATx commands to it. Let me find a modem cable to investigate a little further... Merci !

9.2 up and running for 3 days on my laptop. Big improvements in ACPI support (2.4.22 kernel ?). Bumpy install and some mess in the KDE menus. 270 Mb of updates already. Apart from that, nothing really astonishing. Looks like a well polished 9.1.