Re: Cannot boot 9.1 after OSX installation
On Thu, 2003-11-06 at 12:19, Andreas wrote: Hi all! Yesterday I installed the new OSX Panther on my ibook. Now I cannot boot Linux. My boot partition is on /dev/hda11, swap on 12 and my system on 13. I tried the rescue-system with mount /dev/hda11 /mnt -- dont work, because must specifie the partition type then mount /dev/hda13 /mnt and chroot /mnt ybin -v .. After reboot no bootmanager again. Any ideas? I'm surprised ybin didn't complain about no /proc mounted. Normally you need to mount /proc in the chroot also. You wouldn't normally want to mount the bootstrap, if that's what you're referring to on hda11. -- Stew Benedict -- MandrakeSoft Today I tried mkofboot, no errors but no bootmanager too. After reboot always OSX is starting What had I best do? New installation? No. It seems ybin didn't write to the OF the right command for the boot, so do it yourself: Jump to Open Firmware by pressing COMMAND-OPTION-O-F (apple key + alt key and the O and F key depressed) and reboot. This will bring you the Open Firmware 'shell', then enter the following command: printenv locate the boot-device entry then enter the following command (from memory, basically the last part should be the same as what is shown by the printenv except that the partition number is added (aka 11 in your case)): setenv boot-device hd:11,:\\tbxi Then reboot by entering the mac-shutdown command (or something like that)... Brice
Re: Cannot boot 9.1 after OSX installation
On Thu, 2003-11-06 at 13:19, Andreas wrote: On Thu, 2003-11-06 at 12:19, Andreas wrote: What had I best do? New installation? No. It seems ybin didn't write to the OF the right command for the boot, so do it yourself: Jump to Open Firmware by pressing COMMAND-OPTION-O-F (apple key + alt key and the O and F key depressed) and reboot. This will bring you the Open Firmware 'shell', then enter the following command: printenv locate the boot-device entry then enter the following command (from memory, basically the last part should be the same as what is shown by the printenv except that the partition number is added (aka 11 in your case)): setenv boot-device hd:11,:\\tbxi Then reboot by entering the mac-shutdown command (or something like that)... Won´t work! After reboot now there is a little folder-icon with an question mark on it, than the mac-icon, and than the apple in the middle of the screen for booting OSX. It means that there are no usable yaboot files on your hda11 partition. Are you sure your boot (I mean the apple bootstrap partition, not the partition where vmlinux (usually /boot)) partition is hda11 ? Try the setenv thing and change the partition number... Can you still boot MacOSX ? If yes mount hda11 under MacOSX and have a look to see if it contains the yaboot files... Brice
Re: Cannot boot 9.1 after OSX installation
On Thu, 2003-11-06 at 13:19, Andreas wrote: On Thu, 2003-11-06 at 12:19, Andreas wrote: Hi all! mount /dev/hda13 /mnt and chroot /mnt ybin -v .. After reboot no bootmanager again. I'm surprised ybin didn't complain about no /proc mounted. Normally you need to mount /proc in the chroot also. You wouldn't normally want to mount the bootstrap, if that's what you're referring to on hda11. Yes, Stew is right, you must mount your root partition AND /proc to the chroot if you want to run ybin (yaboot needs to access to the nvram to update the boot-device): boot in rescue from the CD $ mount /dev/hdaXX /mnt where XX is the partition number of / if /usr is on another partition, mount it too $ mount /dev/hdaYY /mnt/usr $ mount -t proc proc /mnt/proc to mount the /proc partition $ chroot mnt /usr/sbin/ybin -v (can't remember exactly where is ybin, I don't have my mac here at work). Send us the output of the last command if it failed. $ umount /mnt/proc $ umount /mnt/usr $ umount /mnt $ reboot I hope this will help you, Brice
Re: odd gurpmi problem
On Fri, 2003-08-22 at 16:23, Isaac wrote: anyone know what this could be? i just get this error and can't install anything via rpmdrake: perl: rpmio_internal.h:447: c2f: Assertion `fd fd-magic == 0x04463138' failed. Aborted i get the same message if i just run gurpmi at the command line. manually installing rpms at the command line works though. I got the same problem a few times, but I can't exactly remember how I recovered from them. Each time it happened because I interrupted a long process of upgrades through urpmi. I guess this is a problem of little-endian/big-endian mismatch somewhere because I never had this issue with x86. Anyway, try to rebuild your rpm database a few times, it might help to solve the problem: $ rpm --rebuilddb Also try to upgrade (by hand) urpmi and its dependencies (perl-URPM, rpm...), and rebuild the database with the new tools. The cooker version I run seems to be immune to this problem... I hope this will help you, Regards, -- Brice
Is cooker PPC dead ?
There seems to be nothing in the ppc subdirs on every mirror I tried, including: ftp://sunsite.uio.no/pub/unix/Linux/Mandrake/Mandrake-devel/cooker/ppc/Mandrake/RPMS As anyone noticed the same problem ? I hope this is a temporary rsync problem... -- Brice
Re: Ibook 600 mdk 9.1 ppc
On Thu, 2003-07-10 at 13:27, [EMAIL PROTECTED] wrote: Yes, I have the same behaviour on my ibook 2rev2 700 combo (MDK9.1). More precisely I boot in text mode (with video=ofonly) and every think is fine but when I start X with [sudo /etc/init.d/dm start], X 4.3 with dri works but the virtual text consoles (alt-ctrl-function keys) aren't useable any longer (the screen is really dark) and it seems to be a contrast problem since the luminosity setting can't fix the problem. There might be a problem with the radeon driver or the screen settings. If I kill X, the console remains dark and I can't use the machine any longuer (except remotely). Does anyone has a clue on how to fix it? Use a recent benh kernel (2.4.21-ben0 or more) to fix that problem. Mandrake Club have a downloadable version. Brice
Re: Ibook 600 mdk 9.1 ppc
On Thu, 2003-07-10 at 15:51, [EMAIL PROTECTED] wrote: Great but: I've found this, but no benh 2.4.21 in rpm: http://rpms.mandrakeclub.com/rpms/MandrakeClub/test/9.1/ppc/kernel-benh-2.4.20-10mdk.ppc.html I usually rsync the sources from the benh repository, then patch for supermount and finally compile it so that the kernel fits my need (aka I remove all the un-needed drivers/features for my ibook). To get the latest 'stable' benh kernel sources: $ mkdir /usr/src/benh_kernel $ rsync -avz rsync.penguinppc.org::linux-2.4-benh /usr/src/benh_kernel (the version you'll get will be 2.4.21-ben2). Happy compilation ! Brice
Re: [Cooker] postfix-2.0.6
On Fri, 2003-03-07 at 00:51, Guillaume Cottenceau wrote: Ok. I was unsure if it was a security fix or not, and since Vincent Danen was not available for immediate answer (timezone difference) and on freshmeat the release target was marked minor security fix, I took the freshmeat stuff for granted. Anyway there were some slight fixes between 2.0.1 and 2.0.6, so it's great to move to 2.0.6 in mdk. http://people.mandrakesoft.com/~gc/pkgs/postfix-2.0.6-1mdk.i586.rpm I'll try it tomorrow. I'll let you know my results. I got the package from a mirror, not directly from your site, and it seems to work pretty well (upgraded from the latest 2.0.1 rpms, the upgrade went smoothly), although I didn't test every possibility... Brice
Re: [Cooker] postfix-2.0.6
`postfix-2.0.6' has a minor security fix. Actually this is not a security fix, but a feature. This is a preventive fix to reduce to 250 characters malformed From: addresses that could affect Sendmail. Postfix by itself is immune to the last Sendmail security problem. Postfix's author decided to add this feature so that Postfix could be installed on a mail gateway in front of Sendmail boxes to protect them, thus leaving time to upgrade those boxes from Sendmail to postfix ;-) http://people.mandrakesoft.com/~gc/pkgs/postfix-2.0.6-1mdk.i586.rpm I'll try it tomorrow. I'll let you know my results. Brice
Re: Opinions on 700MB ISOs?
Any opinions on using 700MB ISOs? I'm OK to have 700MB iso. I don't plan to install my mdk-ppc on old mac... Brice
Re: New benh-mdk kernel build
Hmm, I'll have to look at the mandrake kernel dri and see if it's newer and if it can be wedged into the benh version. Yes, Mandrake version is newer, since it uses the gatos version (I think). And yes you can patch benh's kernel relatively easily, I've done it several times (all the 2.4.19 and some of the 20pre), because mandrake kernel (didn't try the new ones) were not booting (the fb for my radeon was not working at all, slowly transforming the black screen into a green one). But I couldn't have DRI working (I didn't try with the new X version yet) on my 16VRAM ibook (last august model) Brice
Re: [Cooker] postfix, chroot jail, /etc/* updates
On Tue, 2003-01-14 at 08:33, Yves Duret wrote: BTW, how many are interressed of postfix 2.0.1 rpms ? At least me. Brice
Re: [Cooker] postfix, chroot jail, /etc/* updates
On Mon, 2003-01-13 at 12:28, Buchan Milne wrote: Martin Fahrendorf wrote: Am Samstag, 11. Januar 2003 18:12 schrieb Michael Scherer: simply disable the chroot jail. If you ar not permanently connected to the internet and your postfix does not listen to public reachable ports it does not relaly mater. And postfix is known to be realy stable and secure. Plus, one wonders why the default is to chroot, since probably most postfix installations on Mandrake are not production servers, and at present chrooting is more trouble than it is worth for typical use. Postfix as distributed by the author (www.postfix.org) is not chroot-ed by default... Brice
Re: [Cooker] postfix, chroot jail, /etc/* updates
On Sat, 2003-01-11 at 14:19, Peter Lamberg wrote: Just lost dozens of sent mails with postfix during at least 2 months. I use my laptop with postfix in several networks. I have simple scripts to update eg. /etc/resolv.conf. Postfix can't loose mail, since it deletes mail ONLY if the other SMTP server acknowledges it had the mail. If the mail were deferred they still are in the postfix queue (/var/spool/postfix) and still be accessible with the postcat/postsuper commands. mailq (or sendmail -bp) can list the queue. Is there some mechanism that updates /var/spool/postfix/etc/ ? Postfix logs upon restart/reload that files are different, check your /var/log/mail/* log files from time to time. Brice
Re: [Cooker] New nvidia drivers
If you only have 32 megs of memory, you cant run in 1600x1200 24bit color, here is why: memreq = (1600 bit x 1200 bit x 24 bit) / (1024 x 1024) = 44 Megabyte. No: 1600x1200x24 = 4608000 bits / 8 = 576 bytes = 5.5Megabytes If you use 16 bit color: memreq = (1600 bit x 1200 bit x 16 bit) / (1024 x 1024) = 30 Megabyte. 1600x1200x16 = 3072 bits / 8 = 384 bytes = 3.6Megabytes so it should be OK. Brice
Re: Re-installing yaboot from an existing Linux
On Thu, 2002-12-05 at 00:29, Patrick Ladam wrote: My question is: how can I re-install yaboot from my still existing MDK distro? I tried the rescue mode with the install CD to launch ybin but a bunch of binaries used by ybin (hfsutils among others) are not found. Then I edited the ybin script to change the PATH variable to the correct one (change /bin into /mnt/bin and so on) but still commands like 'uname' are not found when running ybin. Boot in rescue mode with the install cd as you already did. Then: $ chroot /mnt Now you can use the already installed binaries (especially hfsutils...) $ ybin It should do it. Brice
Re: sendmail is slow
On Wed, 2002-11-27 at 17:16, Thomas Geenen wrote: On Wednesday 27 November 2002 17:12, you wrote: Le Mercredi 27 Novembre 2002 14:42, Stew Benedict a écrit : BTW - your mailer is overriding the Reply To: for the list so I have to reply to all recipients for my answers to hit the list. This a kmail pb, some poeple said me the same thing (I use kmail). I don't know how to fix this. (my Reply To: is empty on my kmail configuration). i even added my hostname to /etc/hosts so its probably not a lookup issue. Why don't you use postfix ? It is really faster than sendmail, and also really simpler to configure (and I don't talk about the security) the postfix mailing list is helpfull in case of trouble. Postfix is 'compatible' with sendmail, and thus should work fine with kmail. The mandrake package is well done (especially the cooker one), and usually you just need to change the myhostname setting in /etc/postfix/main.cf to have a perfectly running MTA. Beside that, sendmail should log why it is not working properly for you, did you had a look to your mail log (/var/log/mail/*) ? Did you try to send a mail directly with the sendmail interface ? Brice
Re: Migrating to the latest cooker again...
There are 671 packages to install, for about 650 MB. Urpmi give me the following errors for every package: Installing package needs 918Mb on the /var filesystem [snip] I've seen similar messages with the packages with the broken .db files (cyrus-sasl - fixed one on my web space), but the usually reference the / filesystem, and it only effected packages that depended on it. I haven't seen that for /var. It sounds like urpmi is not recognizing the space you've gained with the symlink. I've removed cyrus-sasl from the list of the package to be upgraded and it worked fine ! There must be something wrong with this package. Thank you for your help, Brice
Re: grrr.. yaboot doesn't like me
Any ideas how I can get my cooker install back up and running without doing a reinstall of it as well? Boot to OpenFirmware: 1) switch on, your mac, while pressing Command Option O F. 2) Then enter the magic boot line (which depends on your linux bootstrap partition): setenv hd:[your bootstrap part nb],\\:tbxi And then reboot (can't remember the OF command, something like boot I suppose). Brice
Re: Quest for PPC cooker...
* Desperately no mouse/trackpad. I tried to revert to 8.2 devfsd/hotplug/initscripts/gpm but no chance, and nothing appears in the logs. gpm starts without troubles, but does not log anything. Kdm or X don't log anything unusual about the mouse. What is 'funny' is that booting with devfsd=nomount enable the mouse *but* only with gpm - not with X...so it really seems to be related to /dev entries. About the mouse, again, current cooker initscript fails while loading usb-ohci, but it is OK, since usb-pohci is in the kernel. Ok, then I bet the updated devfsd broke things, as I patched the 8.2 version specically for PPC. The normal mouse setup is: /dev/mouse - /dev/usbmouse - /dev/input/mice I'm not at my machine to give you the fixed /etc/devfsd.conf? lines. Guess what ? You're right ! I found that in fact /dev/usbmouse was linked to /dev/input/mouse0, I changed that to try every other /dev/input entries, and it worked when linking to /dev/input/mice. * No sound - but I'm used to, same trouble with 8.2 - my computer seems rather too new. possibly - don't know offhand on that one dmasound_pmac as included in 8.2 kernel does not work at all. dmasound from benh's kernel (at least rc2, don't know for the latest) works fine - I bet he has an updated version. * Apache/mod_perl complains about missing libperl.so. Once fixed in /etc/ld.so.conf, it complains about the lack of the following symbol perl_init_i18nl10n (is it related to the locale trouble ?) Not sure about that one either, but it doesn't sound necessarily PPC specific (at least the ld.so.conf part). You might drop a line to the maintainer of that package. I think it was related to the locale problem, because when I've updated my locales, the problem disappeared. Anyway, many thanks for your help ! Take care, Brice
Re: Quest for PPC cooker...
Hi all, It seems I'm running into some other problems on my quest for cooker: * the locales-2.3.1.3-1mdk and locales-en-2.3.1.3-1mdk seems broken (even if the md5 sums and the gpg signature are OK): in /usr/share/locale/en/ LC_COLLATE and LC_CTYPE points respectively to ../de_AT.ISO-8859-1/LC_COLLATE and ../br_FR.ISO-8859-1/LC_CTYPE. Since I don't have those two locales, perl keeps complaining about unknown locale. I'm rebuilding the src.rpm to see if it changes something (BTW there seems to be a missing BuildReq in the src.rpm to glibc-i18ndata). * Desperately no mouse/trackpad. I tried to revert to 8.2 devfsd/hotplug/initscripts/gpm but no chance, and nothing appears in the logs. gpm starts without troubles, but does not log anything. Kdm or X don't log anything unusual about the mouse. What is 'funny' is that booting with devfsd=nomount enable the mouse *but* only with gpm - not with X...so it really seems to be related to /dev entries. About the mouse, again, current cooker initscript fails while loading usb-ohci, but it is OK, since usb-pohci is in the kernel. * No sound - but I'm used to, same trouble with 8.2 - my computer seems rather too new. * Apache/mod_perl complains about missing libperl.so. Once fixed in /etc/ld.so.conf, it complains about the lack of the following symbol perl_init_i18nl10n (is it related to the locale trouble ?) Thanks. Brice
Quest for PPC cooker...
Hi all, Since I've bought my iBook, I wanted to upgrade to cooker ppc (as I do it on my pc). My first attempts resulted in a severe crash of the computer (mainly because updating glibc from 2.2.4 to 2.2.5 seems to break the install). I tried again yesterday, but was able to rescue a little bit the computer by booting with the rescue cd and installing again all the rpm in force mode. It crashed about 10 rpms before the end (too bad) with a kind of memory error. It fixes the main problem which was that libgcc1 was not installed (arghh). Anyway after rebooting, my usb mouse or the trackpad are not recognized anymore. It seems to warn about a missing mousedev module. I then booted with the latest benh's kernel, and the mouse's working fine with it. The upgrade didn't change the kernel. Does anyone might have an idea on how to fix this problem ? I was wondering if the proper way to go to cooker ppc wouldn't be to rebuild every src.rpm ? Thanks for your help, Brice
Re: Quest for PPC cooker...
It seems to warn about a missing mousedev module. I then booted with the latest benh's kernel, and the mouse's working fine with it. The upgrade didn't change the kernel. Does anyone might have an idea on how to fix this problem ? Odd - there is no mousedev module, and usb keyboard/mouse are built into the kernel. You're correct the kernel hasn't been updated. It sounds like something just foobarred the links for the mouse, or the usb initscript has issues. Not sure why Ben's kernel would still work though. There was an update of devfsd and initscript - I'll try to downgrade to the 8.2 version. BTW, the new urpmi does not work at all - it seems the perl-URPM package references perl 5.6.1. I was wondering if the proper way to go to cooker ppc wouldn't be to rebuild every src.rpm ? You could I suppose. I don't know why your doing: rpm --rebuild some.src.rpm would be any better than my doing it, but the source is out there. What I just wanted to say is that some rpm are not rebuilt yet on the mirrors (aka perl-URPM for instance), so each time I'm trying to upgrade there are tons of missing dependencies (especially when going to perl 5.8.0), so maybe if I build some missing rpms that didn't go through your compile host yet, I could have more stability (I know that cooker is not a stable release) ? Or am I using a completely broken mirror (ftp.proxad.net listed from mandrake website) ? Anyway, thanks for your help· Brice
Re: Quest for PPC cooker...
Or am I using a completely broken mirror (ftp.proxad.net listed from mandrake website) ? Looking at my repository, I've got /RPMS/perl-URPM-0.50-6mdk.ppc.rpm. A full listing looks like it's up-to-date with perl5.8. (it was built Jul 26). I'm using urpmi regularly myself. its date is Jul 29 on the proxad mirror, and I've updated on sunday (28) so it explains why I've missed it (I may have missed other rpms too) - I'll get it by hand (since my version is totally broken). Everything I've built here should be up on the mirrors by now, I would think. That file is definitely on the mirror source in Paris. At the moment, nothing new for main is coming down the pipe except libvorbis I uploaded today. I'm actually running rebuild on contribs right now. Yes, I've seen some problem with libvorbis, but I personnaly don't care. Aside from everything complaining about a lacking ld-linux.so.2 (which really exists), things seem to be running here OK, with networking, X/kde3, sound, compiler etc. I haven't really spent much time with gnome2, and I know there are a few gnome related packages in my no-build (ie failed) list. That's not to say it's all perfect, but it's functional enough that I'm using it as a build platform. Yes, there was tons of dependence reference to ld-linux.so.2 (which exists how things could work without that ?). If it's useful, I've posted a list of the local RPMs to my web space: people.mandrakesoft.com/~sbenedict/RPMS_Jul_30_2002.txt Cool - Thank you - I'll try to see what are the discrepencies between my versions and yours. I'll keep you posted with my attempts. Thanks for your help, Take care Brice
Re: [Cooker] NEdit 5.3 is out
On Sat, 2002-07-20 at 22:53, cjw wrote: According to www.nedit.org, NEdit version 5.3 was released July 10th. I've just installed the 5.3 package from cooker and there are some strange problem (I got them also with the 5.2 rpm) which does not appear on www.nedit.org pre-built releases: - Immediate crash when using 'split window' menu item - When launching nedit: 'XmFontListCreate() is an obsolete function!' appears on the terminal. - When opening the find window (with CTRL-F) several times for the same file, the search text-edit field is not focused. I mainly uses nedit in server mode, launching only one instance of nedit (and then opening files with ncl). Some of this troubles seem to be related to lesstif (which nedit.org developper have abandonned, they now link with OpenMotif in their builds). Would it be possible to package OpenMotif instead of lesstif and then have nedit link with OpenMotif (since it seems stable, at least more than with lesstif) ? Thanks for your help. Brice
Upgrade to cooker. urpmi troubles.
Hi all, I've succedeed in the install of my 8.2 PPC on my new iBook (radeon enabled). I had to perform the install in text-mode (even video=ofonly did not work). Now everything is working (at least XFree, but no sound, no DRI), and I'm trying to upgrade to the latest cooker, but there's a lot of failed dependencies (all resolvoing to urpmi dependencies). I know that urpmi has changed a lot in the last weeks since I'm also running a cooker on i586. Here is a trace of the problem: libgcc = 3.1.1-0.7mdk is needed by gcc-3.1.1-0.7mdk rpm = 4.0.4-6mdk is needed by perl-URPM-0.06-2mdk perl-base = 2:5.8.0 is needed by perl-MDK-Common-1.0.3-4mdk ... So the question is how is it possible to have such discrepencies ? perl-base (and about all other perl packages) is always 5.6.1 at this time in cooker !! It really seems that the PPC cooker shares somes package with the i586 cooker, but has not been rebuilded/upgraded as often. Thank you for your help. Brice
Re: Upgrade to cooker. urpmi troubles.
On Fri, 2002-07-19 at 12:59, Stew Benedict wrote: [ snip ] Here is a trace of the problem: libgcc = 3.1.1-0.7mdk is needed by gcc-3.1.1-0.7mdk rpm = 4.0.4-6mdk is needed by perl-URPM-0.06-2mdk perl-base = 2:5.8.0 is needed by perl-MDK-Common-1.0.3-4mdk ... Should have had more coffee. Another part of the problem above is noarch packages are put in the PPC tree for me when the i586 build occurs. This is what is driving that 3rd dependency, as perl-MDK-Common is noarch. The other 2 are due to my running out of time before I went on holiday on Jul4 and didn't get everything uploaded I had here. Now I can't upload those missing files, because they are older than the current versions. (libgcc - 0.8mdk on x86). New rpm is in the batch I'll put up today. New gcc still needs to be built. OK, I can now understand the problem. i586 builds containing noarch packages are automatically merged with ppc builds. Seems that it will never be possible for me to have a running cooker ;-( until i586 stabilizes (aka wait for the end of the big move to perl 5.8) Also the machine I build on is not nearly as powerful as the cluster, so packages take up to 10 times longer to compile, even when they do so cleanly, then they have to be uploaded to Paris before they get to the mirrors. I've got about 250 or so to upload today, but you won't be seeing perl. So PPC cooker is moving really slower than i586 cooker. What's the problem with perl under PPC ? Brice
Re: [Cooker] Cannot start MySQL
On Wed, 2002-07-10 at 23:40, Alexander Skwar wrote: So sprach Alexander Skwar am 2002-07-05 um 18:37:42 +0200 : Hmm, there seems to be some problem with libsasl. mysql SIGSEGV's right No ideas? From what I've understood from your strace file: - mysql is dynamically linked with libnss_ldap.so (something in /etc/nsswitch.conf might have triggered the loading of this library). - libnss_ldap.so needs libldap.so (it seems rather normal) - libldap.so needs libsasl.so because LDAP is authenticated using SASL (which is a general authentication scheme, not only bounded to SMTP, even if it is more used with SMTP). - after loading several libs needed both for ldap and sasl (like libcrypto...), mysql tries to find teich.garten.digitalprojects.com IP with the resolver (connecting to /var/run/.nscd_socket). This particular point seems to trigger the reading of /etc/ldap.conf. It then reads the liblogin SASL plugin, and it find it, but is not able to do something with it. The strace report is not sufficient to get exactly what has happened. Maybe you should use ltrace, or gdb directly mysql. What I can see is that something in the ldap.conf is triggering the use of sasl. And that something in nssswitch.conf is triggering the loading of ldap. Send us your nsswitch.conf to see what's the problem. I hope this will help you. Brice
Re: [Cooker] H+/dxr3 in overlay mode
On Sun, 2002-06-23 at 13:01, Borsenkow Andrej wrote: Anybody had success using them in overlay? In which case I am interested in details (video card, resolution, color depth etc). I vaguely recall that not every possible resolution/color depth works but am too lazy to try 'em all :-) I got it to work, my video card is an old AGP ATI Expert@Work in 1152x864 16bits, the sound was not working at all. It took me a couple of hours to get it working fine, because I had to reload the microcode a lot of time (before running xine, sometimes after, don't know why?). Have a look at the FAQ/HOWTO on dxr3.sourceforge.net, and if you follow carefully all the directives you'll have it to work (if your system supports it). Here is what I had to do: * Load the modules (with activate_loopback=1) * Install the microcode (em8300init) * Launch xine with no video window (or whatever supports dxr3 for playing) * run autocal * run dxr3view * play your movie. I hope this will help you. Brice -andrej
[Cooker] multiload-applet-2 keeps crashing.
Hi all, Since I upgraded to cooker last day (and thus moved to Gnome 2), the multiload applet crashes when I add it to my panel (leaving the panel corrupted at the next log-out/log-in). Having a look at gnome's bugzilla ( http://bugzilla.gnome.org/show_bug.cgi?id=77809 ) I found it seems related to the optimisation options when compiling (I have about the same call stack trace than the reporter). Can someone from Mandrake have a look to this problem, or should I recompile this package with less optimisation (I don't know what are the optimisation mdk applies when compiling their packages). Thanks for any help. Brice
Re: [Cooker] multiload-applet-2 keeps crashing.
Hi Installed it a few minutes ago, and it seems to work now. Another question: it seems that the bonobo-activation and the gconf2 servers are not stopped when I log out, is it normal ? Thanks a lot, Brice On Thu, 2002-04-18 at 13:27, Frederic Crozat wrote: On Thu, 18 Apr 2002 12:12:06 +0200, Brice Figureau wrote: Hi all, Since I upgraded to cooker last day (and thus moved to Gnome 2), the multiload applet crashes when I add it to my panel (leaving the panel corrupted at the next log-out/log-in). Having a look at gnome's bugzilla ( http://bugzilla.gnome.org/show_bug.cgi?id=77809 ) I found it seems related to the optimisation options when compiling (I have about the same call stack trace than the reporter). Can someone from Mandrake have a look to this problem, or should I recompile this package with less optimisation (I don't know what are the optimisation mdk applies when compiling their packages). Could you try with gnome-applets 1.99.0 I've uploaded this morning ? I didn't change optimizations but I don't have crashes (but I remember having crashes before).. If it still crashes, please report so I could lower optimizations.. -- Frédéric Crozat MandrakeSoft
Re: [Cooker] Strange BASH problem
Hi Robert, Since yesterday I have the same trouble. So I looked closer to the problem, and surprisingly it comes from pam. Here is what I've found: By strace'ing su I discovered that the time wasted at exit was due to a huge ftruncate call (something was truncating the .xauth/root/your X display file to about 1Gbyte, and then mmap'ing this in RAM) thus driving the kernel crazy. I then further discovered that this was the pam_xauth module that was causing this. So I put it in debug (adding debug in the /etc/pam.d/su file just in front of then pam_xauth line). Then I grabbed the source, and had a look to them. I find that at some point an uninitalized variable (a.size struct member) was used, so its value could be anything. Anyway, here is a patch if someone wants to include it: --- pam_xauth.orig.cFri Feb 1 14:41:52 2002 +++ pam_xauth.c Fri Feb 1 14:44:25 2002 @@ -577,6 +577,7 @@ a.context=Source; a.type=Map; a.level=RdWr|Create; +a.size=-1; /* refcount may need to be up to 1 character larger; may need to a trailing space and cookie if they aren't there to facilitate upgrades from old versions that didn't keep track of the cookie Brice Figureau On Wed, 2002-01-30 at 13:06, Robert Fox wrote: It's not really a problem, but lately (as of the latest Cooker) - when I open a BASH terminal under KDE - then I su to root - do some stuff and type exit, there is an echo of the word exit on the screen and an unusual pause of several seconds before it returns to the normal use account - and when I type exit from the normal windows there is another pause. This does not happen under a normal ASCII screen (Ctrl-Alt-F1) outside of the X server. Strange. R.Fox
Re: [Cooker] evolution hanging in latest cooker
Hi, I encounter the same trouble a few days ago. It seemed that it came from the evolution-executive-summary component. The only solution we found was to edit the '/usr/bin/killev' script, extract the executive-summary part, launch evolution (which will block forever), then kill the executive summary, and then voila... Maybe evolution was stuck in trying to access a website somewhere to get its RSS/RDF file for the executive. I wasn't able to find where are stored the executive summary configuration, so I couldn't change them to 'more secure' ones. Hope this helps. Brice On Tue, 2002-01-29 at 22:31, Brian J. Murrell wrote: I have cooker updated as of this morning, including glibc and kernel and I seem to be having an issue with evolution. It starts up mostly normal like but just before painting the main widget it hangs. I have tried removing ~/evolution and also rebooting and the problem still seems to persist. Any ideas? b. -- Brian J. Murrell
Re: [Cooker] Re: evolution hanging in latest cooker
Edit /usr/bin/killev and put $debug=1 at the beginning. Run it (without evo running otherwise it'll kill it), then grab the first four lines (which are related to executive-summary), then launch evo, wait for it to block. Here are the four kills for my system (it may depends of the system you're running because killev uses oaf to retrieve the components names...): killall -9 evolution-executive-summary killall -9 lt-evolution-executive-summary killall -9 evolution-execut killall -9 lt-evolution-exe Then paste them in a shell. Should destroy the bad component. Brice On Wed, 2002-01-30 at 11:31, Brian J. Murrell wrote: On Wed, Jan 30, 2002 at 09:57:47AM +0100, Brice Figureau wrote: Hi, Hello, I encounter the same trouble a few days ago. It seemed that it came from the evolution-executive-summary component. That would look right here too. The only solution we found was to edit the '/usr/bin/killev' script, extract the executive-summary part, Which was which part. I experimented but could not isolate it. launch evolution (which will block forever), then kill the executive summary, and then voila... From the testing I did do this looks like it will work. I just need to know how to isolate the exec-summary part. Maybe evolution was stuck in trying to access a website somewhere to get its RSS/RDF file for the executive. That is what I was thinking. Bad design if so. Hope this helps. It does, thanx. b. -- Brian J. Murrell