Re: [Cooker] NVIDIA Driver help please!!
Eugenio Diaz wrote: Mark Finlay wrote: I'm running 8.2 with the cooker kernel cus the 8.2 kernel only runs my via chipset as ata2. but now i cant compile the NVIDIA kernel driver cus the cooker kernel was compile with gcc3.1 AKAICT. Would someone do me a big favour and save me 36.6ing gcc3.1, and do a rpm --rebuild on the NVIDIA kernel src.rpm and send me the resulting RPM. The src.rpm is here http://download.nvidia.com/XFree86_40/1.0-2960/NVIDIA_kernel-1.0-2960.src.rpm CC: me Pretty Please :) Just do this before compiling export IGNORE_CC_MISMATCH=1 Or better yet, compile it with the same version of gcc that was used to compile the kernel [dmc@roc dmc]$ cat /proc/version Linux version 2.4.18-6mdk ([EMAIL PROTECTED]) (gcc version 2.96 2731 (Mandrake Linux 8.2 2.96-0.76mdk)) #1 Fri Mar 15 02:59:08 CET 2002 Then do something like export CC=gcc-2.96 instead of the above. Obviously you must install the gcc-2.96 packages(rpms), which can happily live alongside whatever other versions of gcc/egcs you have installed. Note, I haven't actually tried the above line, but either it, or something substantially similar will work for you, and I'm pretty sure it will be a much more robust solution than compiling the module with a different version of gcc than was used to compile your kernel. -dmc
Re: [Cooker] NVIDIA Driver help please!!
Borsenkow Andrej wrote: I'm running 8.2 with the cooker kernel ^^ Or better yet, compile it with the same version of gcc that was used to compile the kernel [dmc@roc dmc]$ cat /proc/version Linux version 2.4.18-6mdk Have you bothered to actually read the question? -andrej No. :) -dmc
Re: [Cooker] Emu10k1 Mixer settings
Nathan A. Smith wrote: Hi, I seem to be having trouble getting my sblive soundcard working properly. I have sound (system, xmms, etc) - but only from my rear speakers I have often run into people who had just their rear speaker output connected to their normal primary speakers under windows, then being perplexed when they got no sound in linux. I suggest you double check and make sure that what you think is the front speaker output jack, really is the front speaker output jack. -dmc and I can't get my tv sound working (I have verified the hardware settings by trying it in windows). I have tried alsamixer and I can adjust treble/bass/wave sur but nothing else does anything. To make matters worse it hard to tell what is suppose to do what. Is there some place I can look for what mixer settings do what, helps us (not so bright) people configure our soundcards to play tv sounds and have thier speakers all work?? Thanks in advance Nasa System: Athlon 500, 512Megs memory, ATI 8500DV, sblive, and some other stuff.
Re: [Cooker] init functions fixes
Liam R. E. Quin wrote: the security stuff is to do with unquoted shell variables Can you explain, or give me a pointer to a relevent faq/document? I found the NCSA Secure Programming Guidelines, and it mentions the IFS thing, but nothing about quoted vs unquoted variables. It also fails to mention why setting IFS is a good thing. Really I'm just looking to develop good shell scripting style. So if their are guidelines on when you should quote things, and when not to (even when both ways seem to work), I would like to know. As well as any other similar techniques. -dmc - probably IFS should be set somewhere, too, for the case where someone does su from a malicious user's terminal, then runs an init script. hmm, evolution crashes if I attach a file, I'll paste it... and if that fails, back to mutt :-) Liam
Re: [Cooker] re : multimedia keyboards
Liam R. E. Quin wrote: On Mon, 2002-05-27 at 11:17, Doug McClendon wrote: xbindkeys works great for me. How does this compare to xkeycaps? Its been awhile since I've even seen xkeycaps, did it get removed from mandrake? But from memory, xkeycaps is something completely different. If I recall, its a tool for mapping lower level keycodes to X level keycodes, allowing one to do things like switch the ctrl and caplocks keys. xbindkeys is like hotkeys, khotkeys, lineak, in that it is a daemon which has a configuration list of X level keycodes, and a shell command line per keycode. Then, whenever one of those keys gets hit, corresponding command line gets executed. -dmc
Re: [Cooker] re : multimedia keyboards
xbindkeys works great for me. I think its a better solution just on the basis that its versatile, and isn't limited to the 'extra' keys. I.e. you can bind arbitrary commands to the multimedia keys, and ctrl+alt+x, etc It lacks cute knowlege of the different keyboard types and enumerations of their special keys. But it works just great if you create a new entry, click capture, hit your special key of choice, name it, then specify a command line. It is also a fairly mature package (been cranking out new releases regularly for over a year). The only downside I've noticed, is that compared to the hotkeys(tm) package (which is also 'mostly' limited to multimedia keys), that xbindkeys does not work when X is in fullscreen DGA mode (xmame fullscreen). Somehow hotkeys(tm) worked for me even then. -dmc Guillaume Cottenceau wrote: Cosmic Flo [EMAIL PROTECTED] writes: Finally, it's not a bug : Here a message from a developer of the project team : Unfortunately, lineakconfig 0.1 is not compatible with lineakd 0.3, due to different file format... as the release notes already state. lineakconfig 0.2 is in development and will be compatible with lineakd 0.3 They are stupid.. their package `lineakconfig' needs `lineakd' to work, but latest version of `lineakd' doesn't work with latest version of `lineakconfig', and no configure script states so, same for program startup.. In the mean time, either use lineakconfig in conjunction with lineakd 0.2 or use lineak 0.3 but configure the keys with a text editor. Time will flow until I have enough motivation to test again their program :-). Up to you to package a working version anyway.
Re: [Cooker] Hanging for the cdrom ...
Stéphane Teletchéa wrote: Dear sirs, i would agree if the errors appeared only when i pass ide-scsi at boot time. But when i didn't pass it, i can neither access the cdrom drive !!! When i use the linus kernel, i have the same problem, so i think it is a kernel-related problem, not just llinked to ide-cd nor ide-scsi nor supermount. Stef Can you access the cdrom drive when the same computer is booted into another OS? -dmc
Re: [Cooker] Hanging for the cdrom ...
Seth Zirin wrote: This appears to me to be neither a kernel problem nor a cooker problem - just a simple configuration error. While the equivalent of hdc=ide-scsi must be appended to the boot command for a CD-writer in this scenario, you must also arrange for the ide-cd driver to ignore the device by adding something like options ide-cd ignore=hdc to /etc/modules.conf. The ide-scsi driver cannot control a device once it is claimed by the ide-cd driver. Thats not exactly true. Ever try something like this in an init script- (if someone wants to offer a cleaner way to do this, I'm all ears) pushd /proc/ide /dev/null for ide_interface in $( echo ide? ); do [ $ide_interface = ide? ] continue cd $ide_interface for ide_device in $( echo hd? ); do [ $ide_device = hd? ] continue cd $ide_device if ( cat model | grep -q ^CD-R ); then echo -n ide_scsi:1 /proc/ide/${ide_interface}/${ide_device}/settings echo -n ide-scsi /proc/ide/${ide_interface}/${ide_device}/driver fi cd .. done cd .. done popd /dev/null # load ide-scsi and sr_mod modules now that any cd-writers have been # told that they will be using the ide-scsi+sr_mod driver /sbin/insmod -q ide-scsi /sbin/insmod -q sr_mod The success with other kernels is likely caused by kernel configuration differences. The ide-cd driver will always ignore a CD-Writer when the driver is not configured. Further information is available in the CD-Writing HOWTO... [EMAIL PROTECTED] -Original Message- From: Stéphane Teletchéa [mailto:[EMAIL PROTECTED]] Sent: Wednesday, May 22, 2002 6:16 AM To: [EMAIL PROTECTED]; Ron Stodden Subject: Re: [Cooker] Hanging for the cdrom ... Le Mercredi 22 Mai 2002 14:00, Stéphane Teletchéa a écrit : Le Mercredi 22 Mai 2002 09:45, Stéphane Teletchéa a écrit : Le Mercredi 22 Mai 2002 04:30, Ron Stodden a écrit : Stéphane Teletchéa wrote: Le Mardi 21 Mai 2002 13:47, Ron Stodden a écrit : OK, thanks. You should have idc=ide-scsi in the append of _all_ linux stanzas. This and issue lilo as superuser and reboot is all that is needed to make your cdwriter accessible to reading as /dev/scd0 or /dev/sr0 (mount in your /etc/fstab on /mnt/scd0 (you must mkdir this directory yourself), for example) and for writing as the dev=n1,n2.n3 given by cdrecord -scanbus. If it was as easy as this ... It is. Mind you, I do not use any supermount, nor see any need for it. Try it without supermount - if OK, complain about supermount. I tried without and with supermount, this is the first thing i did, but it didn't change anything. That's why i tried different kernels. That's why i told you it would be far too easy. Stef I'm using for the moment 2.4.3-20mdk kernel which works perfectly for the cdrom drive. Stef As told me wally, i used the 2.4-linus kernel and the problem is still there, so it is definitely a kernel problem. I report them my complain. Happy coding. Stef
[Cooker] [Fwd: [expert] cannot recompile kernel on 82 (rpm --rebuild)]
Well, I tried posting this on [expert], since its an 8.2 question and not a cooker question, but I'm not finding a wealth of help, or even of basic understanding over there. Specifically, on mdk8.2, an rpm --rebuild of the kernel .src.rpm is failing in the middle of the kernel compile. Can anyone help me out? Attached is the original post. Just to avoid any stupid responses like I got on [expert], yes, I do in fact know what I'm doing, and do in fact have a good reason for doing it. I'm assuming that somewhere, somehow, someone did an rpm --rebuild of the kernel .src.rpm to generate the kernel rpms that are in mdk82, I merely want to recreate that, so that afterword I can add my own custom kernel patch, to be used in a custom distribution. -dmc ---BeginMessage--- I tried rpm --rebuild /home/dmc/kernel-2.4.18.6mdk-1-1mdk.src.rpm and after half an hour, got this /usr/bin/gcc -D__KERNEL__ -I/usr/src/RPM/BUILD/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -DMODVERSIONS -include /usr/src/RPM/BUILD/linux/include/linux/modversions.h -DKBUILD_BASENAME=cache -c -o cache.o cache.c /usr/bin/gcc -D__KERNEL__ -I/usr/src/RPM/BUILD/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -DMODVERSIONS -include /usr/src/RPM/BUILD/linux/include/linux/modversions.h -DKBUILD_BASENAME=journal -c -o journal.o journal.c In file included from /usr/src/RPM/BUILD/linux/include/asm/semaphore.h:39, from /usr/src/RPM/BUILD/linux/include/linux/fs.h:215, from journal.c:11: /usr/src/RPM/BUILD/linux/include/asm/system.h:203: parse error before `volatime' /usr/src/RPM/BUILD/linux/include/asm/system.h:204: warning: function declaration isn't a prototype /usr/src/RPM/BUILD/linux/include/asm/system.h: In function `__xchg': /usr/src/RPM/BUILD/linux/include/asm/system.h:205: `size' undeclared (first use in this function) /usr/src/RPM/BUILD/linux/include/asm/system.h:205: (Each undeclared identifier is reported only once ... ... ... Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com ---End Message---
Re: [Cooker] maximal install issue (self fixed automake rebuild xmmsproblem)
Guillaume Cottenceau wrote: Doug McClendon [EMAIL PROTECTED] writes: When you tried to remove automake1.5, what other package got removed? I wonder why did you end with both versions of automake, which conflict... I just did rpm -e automake1.5. I don't recall it complained about any dependencies breaking. :-((. Can you send me /root/drakx/report.bug.gz ? on its way Just complaining... Mandrake still rocks. But it would be very nice if there was a dedicated 'maximal' install path, as there is a dedicated 'minimal' install path. I mean really, 4G is nothing nowadays :) Preferably also a path that would prevent - it has no real meaning, especially when there is a meaningful package group selection during install, and that urpmi and rpmdrake do work to install missing stuff Well, then perhaps its a bug in the 'meaningful package group selection', because I selected *everything*, then did individual package selection, and selected *everything*, and double checked 2 or 3 times making sure that every last checkbox was checked. And yet, these 3, and other rpms did not get installed. The groups mean the main, most important packages, handled by groups. These packages can be also found in the default package selection list, also grouped. If you want to see all the available packages, you need to toggle the flat ordering of the package selection. So, you are in the 'individual package selection', and you can choose between either 'flat list', and 'tree list'. Are you saying that the tree list doesn't include everything thats in the flat list? I had just assumed the flat list un-treeified them so you could find things alphabetically. This is another story behind my desire for a 'one button' maximal install path. My intuitive guess was that if I just expanded the entire tree of individual package selection, and checked everything, that I would get everything. hdparm, libMesaglut3-devel, gphone I can't tell you whether or not checkboxes existed for these items, but I can tell you that I had every last checkbox selected. I can tell you that I just checked my /root/drakx/auto_inst.cfg.pl and install.log, and there is no mention of them. The reason I asked for a maximal install path, is because I want there to be one button I push, that installs absolutely everything, except in the case where there is direct conflict, in which case I want the more stable choice selected.
Re: [Cooker] Re: installing live into another device
Pixel wrote: You can try using urpmi --root /the_dir ... Wow, urpmi really does have --root, documentation (other than src) not withstanding. Maybe now I can stop using chroot. --excludedocs (and generic pass thru to rpm) might be a nice next feature for urpmi. -dmc
[Cooker] maximal install issue (self fixed automake rebuild xmms problem)
I was about to post a problem rebuilding xmms from src.rpm, but I fixed it myself, and just figured I should post anyway. My complaint/issue/problem is that I installed mdk (8.2) as 'maximally' as the install allowed. I just selected every last package box in the install. First, I was very disappointed when several packages that I normally use, didn't get installed this way cough cough hdparm, xosd, libMesaglug3-devel, gphone, etc.. etc.. etc... Now, I discovered that both automake-1.4 and automake1.5 were installed, and it was causing my 'rpm --rebuild xmms-blabla.src.rpm' to fail, until I took 1.5 out. I stared in awe for awhile when I did 'rpm -q automake' and it returned -1.4, when doing 'automake --version' returned 1.5. Just complaining... Mandrake still rocks. But it would be very nice if there was a dedicated 'maximal' install path, as there is a dedicated 'minimal' install path. I mean really, 4G is nothing nowadays :) Preferably also a path that would prevent me from being able to install both versions of automake simultaneously (maybe thats just a bug though). -dmc
[Cooker] 8.2: autofs starts portmap even when portmap is off
in 8.2, /etc/rc.d/init.d/autofs has this nasty habit of starting portmap even if you have done chkconfig --del pormap. Not a nice feature IMO... -dmc
[Cooker] $PROMPT in rc.sysinit, how to set?
Is there a 'proper' way to set $PROMPT=no in rc.sysinit. Some sort of option file, or command line arg? Something other than manually putting it at the top of rc.sysinit? I'm going for an ultra quite boot using the kernel arguments console=tts/2 quiet but I still see (in RC1) the Press I for... spewed to the default screen (/dev/vc/1 I think). Reading rc.sysinit, the PROMPT=no appears to be a simple way to disable that (and other stuff I don't want anyway), but I don't see by what mechanism (other than manually coding it in the file) it gets set. -dmc
Re: [Cooker] is there a way to link hdc to ide-scsi after boot?
SI Reasoning wrote: I just modularized ide-cd floppy. I have to say... I can't tell the difference in use of the dvd/cdrom drive between ide-cd and ide-scsi. I had to change my fstab to /dev/cdrom0 (for /dev/hdb) and that was the only difference. For the most part I can't tell any difference, except- The main reason I care about this stuff is that I am playing with generating bootable cdroms. And some time ago I tried the simple path of just not using ide-cd at all, and it _seemed_ to me that my bootable cdrom started failing more often (with relatively old atapi-cdrom drives). I didn't exactly meticulously prove that it was the fault of the ide-scsi+sr_mod vs ide-cd, but I just went on the gut feeling that ide-cd is probably way more tested and polished than ide-scsi+sr_mod. So I went down the path of modularizing, and using ide-cd to boot, then being able to remove it and use ide-scsi in case I wanted to use a cdburner. Since I have an older cooker, those changes may have already been made. Ide-cd is loaded by default, how can I change that? I posted a request about a week ago that ide-cd be modularized by default. For the masses it probably requires much more thought that I'm willing to give it. But I'm just letting people know that its possible, especially since that append=hdd=ide-scsi stuff seems like the ugliest kludge in the world, for something as common as a cdburner. -dmc
Re: [Cooker] is there a way to link hdc to ide-scsi after boot?
Borsenkow Andrej wrote: ÷ óÒÄ, 13.03.2002, × 08:48, Doug McClendon ÎÁÐÉÓÁÌ: Why yes! The only way I have found so far is to compile ide-cd as a module. The problem (without kernel arguments 'hdx=ide-scsi') is that if the ide-cd driver is compiled in, then it will claim the device during boot/initialization, and AFAICT there is no way to tell the idecd driver to release the device so that the ide-scsi driver can take it. (the kernel argument is what prevents the ide-cd driver from claiming it at boot time) If however, ide-cd is a module, then it can be unloaded (or not loaded in the first place), after which loading ide-scsi will successfully take over the device. Have you actually tried it? yes When ide-scsi is loaded it takes over only those devices that declare they are using ide-scsi driver. And the only known to me way to set hdc to use ide-scsi driver is to pass hdc=ide-scsi on boot. I am pretty sure you are mistaken. My understanding is that the cmd line args (hdc=ide-scsi) are for the benefit of the _ide-cd_ driver, to tell it to back off of that device. It is not an argument to the ide-scsi driver telling it what to use. Both drivers simply take over all devices that they were not specifically told to stay away from. There are only two ways I know to prevent the ide-cd driver from taking over a device, one is to tell it to stay away from it via a cmd line arg, the other is to have it be a module and remove it from the face of the kernel. ide-scsi driver is horrible hack. It badly needs rewrite. But I am always scared by the fact that it touches SCSI subsystem that is anything but easy to ubderstand. Ideally ide-scsi should be on top of IDE not replace original IDE driver. May be when I have enough courage :-) This confirms my anecdotal experience, which is why for my bootable cdrom, I load the ide-cd driver from initrd for the general purpose of booting, and then unload it and load ide-scsi in the case that I want to use the cdburner. This much is hard fact, I use my cdburner and ide-cd without any ugly (hard coded) kernel command line args. But so far as I know, there is no way to do it if ide-cd is compiled into the kernel. -dmc
Re: [Cooker] rc1 urpmi --auto pciutils fails in an odd way...
Nope, upon reproduction, it worked correctly (installed just pciutils, made no mention of -devel). I'll send you /var/lib/urpmi/synthesis.* if I ever see it again. -dmc François Pons wrote: Doug McClendon [EMAIL PROTECTED] writes: [root@roc /]# urpmi --auto pciutils installing /urpmi_media/mdk82/Mandrake/RPMS2/pciutils-devel-2.1.9-2mdk.i586.rpm error: failed dependencies: pciutils = 2.1.9 is needed by pciutils-devel-2.1.9-2mdk Installation failed Is it reproducible currently, if yes, can you send me synthesis file of all your media ? Thanks, François.
Re: Brute force method to reset ide-scsi for a CD on vanilla Mandrakekenel Re: [Cooker] is there a way to link hdc to ide-scsi after boot?
Guillaume Cottenceau wrote: Borsenkow Andrej [EMAIL PROTECTED] writes: seems possible. Anyway, here is brute force method to force ide-scsi after boot: echo -n ide_scsi:1 /proc/ide/hdc/settings rmmod ide-scsi echo -n ide-scsi /proc/ide/hdc/driver modprobe ide-scsi And it even -works-! congrats Andrej! Maybe we could even wrap cdrecord with that... (and we can even roll back to IDE support, kernel is sometimes positively surprising ;p) Congrats... Much better than my modularization method. Thanks a ton!!! But why do you call it brute force other than just the right way? I'm assuming that the proc interface that it uses, is designed for just such a reason, no? -dmc
[Cooker] rc1 urpmi --auto pciutils fails in an odd way...
[root@roc /]# urpmi --auto pciutils installing /urpmi_media/mdk82/Mandrake/RPMS2/pciutils-devel-2.1.9-2mdk.i586.rpm error: failed dependencies: pciutils = 2.1.9 is needed by pciutils-devel-2.1.9-2mdk Installation failed
Re: [Cooker] is there a way to link hdc to ide-scsi after boot?
Why yes! The only way I have found so far is to compile ide-cd as a module. The problem (without kernel arguments 'hdx=ide-scsi') is that if the ide-cd driver is compiled in, then it will claim the device during boot/initialization, and AFAICT there is no way to tell the idecd driver to release the device so that the ide-scsi driver can take it. (the kernel argument is what prevents the ide-cd driver from claiming it at boot time) If however, ide-cd is a module, then it can be unloaded (or not loaded in the first place), after which loading ide-scsi will successfully take over the device. Right now this would require you to recompile your kernel with CONFIG_BLK_DEV_IDECD=m Which is probably more hassle than you are looking for, but if you really want it as I do, thats the best solution I have found so far... -dmc SI Reasoning wrote: I was wondering if there was a way to link hdc to ide-scsi after boot (instead of append=hdc=ide-scsi in kernel) = SI Reasoning [EMAIL PROTECTED] A requirement of creativity is that it contributes to change. Creativity keeps the creator alive. -FRANK HERBERT, unpublished notes __ Do You Yahoo!? Try FREE Yahoo! Mail - the world's greatest free email! http://mail.yahoo.com/
Re: [Cooker] is there a way to link hdc to ide-scsi after boot?
SI Reasoning wrote: How would you handle 1 cdrom/dvd ide-cd and 1 cdrw ide-scsi after boot if I want both? Good question :) I haven't exhaustively looked through the docs/source to find out if and how it is possible to have both modules loaded, telling each module to look at different devices. *Maybe* its something as easy as modprobe ide-cd hdd=ide-scsi modprobe ide-scsi hdd would be used by ide-scsi, all the rest by ide-cd Since the driver recognizes the arg on the kernel comline, maybe it works there as well. If that doesn't work (and in fact is an absent ability), one option is to let the cdrom/dvd be handled by the ide-scsi driver. If you load sr_mod (the scsi cdrom module) after the ide-scsi module, you should see all of your ide cdrom/cdrw/dvd drives come up as scsi cdrom drives (old /dev/scdX, now /dev/cdroms/cdromX linked to I forget where ../scd). The only downside to this is that the ide-scsi+sr_mod driver may be less efficient or less reliable than the desired ide-cd driver (just guessing). -dmc
Re: [Cooker] request: ide-cd compiled as module
Pros: You can on the fly decide whether you want your cd burner to be recognized by the ide-cd driver, or the ide-scsi driver. * no more kernel command line args needed for cdburner to work * Afaik the kernel needs the boot option to recognize it as scsi. But what exactyly would be the pro to be able to switch between ide or scsi? The kernel needs the option because ide-cd is initialized at boot (not during module load). Therefore it will 'claim' the cdburner for itself unless preempted by a command line argument. If however ide-cd is a module, it can either be not loaded, or rmmod'd, and the subsequent loading of ide-scsi can claim the cdburner without the need of hdx=ide-scsi. I haven't looked into the issue comprehensively, such as how with modules, to get one device working with one driver, and another with the other. I did try for a while just nuking the ide-cd driver, but it seemed like the ide-scsi(and sr_mod) did not work as well across the board with all cdroms as well as the ide-cd driver. -dmc
Re: [Cooker] new nvidia drivers
J.P. Pasnak wrote: On Thursday 07 March 2002 09:56, you wrote: If you are interested in evil proprietary software :-) Since I am at work, haven't got the time to try on cooker yet. Release Highlights for 1.0-2802: * Anisotropic filtering support ^^what the hell is that? Danny I put up a couple of links to the explanation of anisotropic this morning: http://www.warpedsystems.sk.ca/article.php?sid=521 Are you sure that explanation is the relevent one? I suspect this explanation is the more relevent one... http://oss.sgi.com/projects/ogl-sample/registry/EXT/texture_filter_anisotropic.txt (The two don't seem to be very similar) -dmc
[Cooker] request: ide-cd compiled as module
A request: That ide-cd be compiled as a module Pros: You can on the fly decide whether you want your cd burner to be recognized by the ide-cd driver, or the ide-scsi driver. * no more kernel command line args needed for cdburner to work * Cons: For some occasions you might need to have the module in the initrd and load it via nash. If thats the only con, I really don't see any reason not to do it. I suppose its a mute point long down the road if the all drivers will be modules thing comes to pass. -dmc
[Cooker] where can I find srpms for beta3?
Where can I find the srpms for the beta releases? Presumably this equates to where can I find an archive of cooker srpms? No, I don't mean the latest srpms, I mean the ones that built the beta(3) binaries -dmc
[Cooker] does bootsplash not work with 640x480?
I'll probably just try this, but does bootsplash not work for 640x480? The docs specifically mention 800x600, 1024x768, and 1280x1024, but leave out 640x480 What gives? Personally I think 640x480 is a wonderful video mode :) -dmc
Re: [Cooker] Sound problems in 8.2betas
What output are you using? If spdif, then perhaps emu-config --digital or emu-config --analog are what you are looking for. The symptoms of the wrong setting would be as you described (everything looks ok, but silence). (from emu10k1-tools*rpm) -dmc [EMAIL PROTECTED] wrote: On Wed, 27 Feb, at 23:06:26 +0800, Leon Brooks [EMAIL PROTECTED] done said: On Wednesday 27 February 2002 22:41, [EMAIL PROTECTED] wrote: On Wed, 27 Feb, at 09:37:23 +0100, Stefan Siegel [EMAIL PROTECTED] done said: Start aumix to see if the master volume is set to 0. Yeah, that was my first thought too. I turned everything up as high as it could go, but still nothing. All unmuted? Used alsamixer for an ALSA system? Yep, I turned everything up full blash and still no love. I fired up alsamixer and turned everything up there too, but nada. Quite frustrating, -Charlie
Re: [Cooker] mdk82b2: urpmi is lying to me!
Borsenkow Andrej wrote: On Пнд, 2002-02-25 at 12:54, François Pons wrote: Doug McClendon [EMAIL PROTECTED] writes: urpmi seems to be lying to me. It says its going to add kernel-source, but instead it adds all the dependencies, and then silently drops the main rpm I asked for... There is a problem in kernel source package, remove kernel-source in file /etc/urpmi/inst.list as urpmi didn't manage package to install instead of upgrade nicely. Sorry? I have been installing kernel-source with urpmi all the time. May be the original poster meant kernel SRPM? Then urpmi obviously does the right thing. It was not the SRPM, but the one from the mdk82b2 iso image. It also was an install, not an upgrade. Path looks something like 8.2b2 ultraminimal install install 9 rpms for urpmi (urpmi is 9th) init urpmi pointing it at at basemedia (RPMS2/* from cd2 copied to RPMS directory) urpmi --auto kernel-source # kernel-source NOT installed, but prereqs are installed urpmi --auto kernel-source # kernel-source IS installed -dmc
[Cooker] mdk82b2: urpmi is lying to me!
urpmi seems to be lying to me. It says its going to add kernel-source, but instead it adds all the dependencies, and then silently drops the main rpm I asked for... In my output below, notice how I rerun the same command line again, and it actually installs it the second time. (make postinstall script is ??? another failure like one I just posted about wget, although this one is a bit different as presumably urpmi is not passing --excludedocs to rpm ) [root@roc /]# urpmi -v --auto kernel-source read synthesis file [/var/lib/urpmi/synthesis.hdlist.mandrake_82_basemedia.cz] no entries relocated in depslist installing /urpmi_media/mdk82/Mandrake/RPMS/kernel-source-2.4.17-16mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/kernel-headers-2.4.17-24mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/libncurses5-devel-5.2-19mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/libbinutils2-2.11.92.0.12-6mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/make-3.79.1-6mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/gcc-2.96-0.74mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/glibc-devel-2.2.4-24mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/binutils-2.11.92.0.12-6mdk.i586.rpm /urpmi_media/mdk82/Mandrake/RPMS/gcc-cpp-2.96-0.74mdk.i586.rpm starting installing packages error: failed dependencies: glibc-devel is needed by kernel-source-2.4.17-16mdk ncurses-devel is needed by kernel-source-2.4.17-16mdk make is needed by kernel-source-2.4.17-16mdk gcc is needed by kernel-source-2.4.17-16mdk Preparing... ## kernel-headers ## libncurses5-devel ## libbinutils2 ## make ## install-info: No such file or directory for /usr/share/info/make.info.bz2 error: execution of %post scriptlet from make-3.79.1-6mdk failed, exit status 1 glibc-devel ## binutils ## gcc-cpp ## gcc ## [root@roc /]# ls /usr/src/ RPM/ [root@roc /]# urpmi -v --auto kernel-source read synthesis file [/var/lib/urpmi/synthesis.hdlist.mandrake_82_basemedia.cz] no entries relocated in depslist installing /urpmi_media/mdk82/Mandrake/RPMS/kernel-source-2.4.17-16mdk.i586.rpm starting installing packages Preparing... ## kernel-source ## [root@roc /]#
[Cooker] 8.2b2, wget postinstall script problem in conjunction with --excludedocs
Very minor bug, when doing rpm -i of wget on mdk82beta2 with the --excludedocs option, the postinstall script fails because it is trying to act on one of the documentation (info) files that didn't get installed. -dmc
[Cooker] mdk82b2: misc rpm dependency issues. perl-SDL doesn't require perl?
I'm trying to build minimal installations. I took the minimal of minimal installations, then installed urpmi and its 9 needed rpms. Then I used urpmi to add frozen-bubble, which also installed perl-SDL. frozen-bubble uses perl scripts. Then I tried removing in reverse order urpmi and its 9 dependencies, one of which is perl. I was dismayed to find that it let me remove perl. It seems that perl-SDL doesn't depend on perl, nor does frozen-bubble. (even though they clearly should) I also noted that I could urpmi xmms, and it wouldn't install XFree86-server. Clearly what I'm doing can be considered corner-case, but hey... I am running these things non native, i.e. rpm --root=, and chrooting urpmi. As well as other evil things that probably no one else will ever do. But from what I can tell, the dependencies could use some help. -dmc
[Cooker] bug(?) with default modules.dep in mdk82b2
My problem? /lib/modules/2.4.17-16mdk/modules.dep from the kernel rpm in mdk82b2 references .o files while said rpm has .o.gz files. Before you ask me why in the hell could I possibly care, tell me why in the hell it should be the way it is instead of the other way. And in an unrelated but related matter, is there any already used (as opposed to theoretical) way to install mandrake to a partition (or looped fs, or just a dir) from a normal running system, as opposed to a system booted from installation media? What I'm searching for is a script (or a way to write a script) that would look something like this gen_mdk_inst auto_replay_floppy.img destination_directory -dmc