adamw

for whatever you're trying to compile - look at the name of the lib that it says is missing, and see if there's a similarly named -devel package you can install.

to clarify chris's response, the "install not upgrade" is more of an information line than a warning. it's normal to see it. Maybe your RPM database is FUBARed? Try this, as root: rm -f /var/lib/rpm/__* rpm --rebuilddb

you should have found that either in the description or the documentation for the PLF package it tells you where to put the necessary quake3 files.

btw, I don't like the look of the app that started this thread. it's non-free and has an ugly non-standard interface. no thanks.

hmm, I think I did that. not 100% sure, actually. I'll have to test it out tonight...

make sure you have a .gtkrc-2.0 in /root and all users' /home directories.

bvc: RPMS2 is just a symlink to contrib, on mirrors. I think RPMS3 points at jpackage.

mplayers -channels option may help - check the manpage. I'm not entirely sure what passing -channels 1 does, but can't hurt to try. (I just use -channels 6, for 5.1 :>)

btw, I thought yahoo shut off their free POP3? there's a script called fetchyahoo or something which is designed to pull mail to an mbox or maildir through the web interface...

definitely won't be the problem with 10.0, as it uses a much earlier kernel. Shouldn't be the problem with 10.1, either. Can you give more detail? dmesg would be a good start.

fresh install will always do some things urpmi upgrade cannot / does not do. urpmi won't ever install a completely newly created package that isn't required by a package you already have, for instance, and new drak tools often fall through this net. It also won't update configuration files you edit yourself, which can lead to new features not appearing, or (occasionally) problems with updated applications. urpmi upgrading generally does an excellent job, though, especially if you do it on a continual basis.

chkconfig is nice and fast for when you don't need the frontend, too. just "chkconfig service off" does the trick.

ixthusdan, linux viruses / worms do exist. they're generally known as rootkits, as their aim is usually to get access to the root account. they are quite rare, though, which leads me to suspect that maybe that anti-virus thing made a false detection and deleted something vital...

it was more of a technical note than a recommendation :). and the trademark is for Exciting Things, not the product...hehe

Well, for 10.0 the procedure was to use the CE tree (although it didn't matter as CE and Cooker were literally identical). I'd *expect* it to be the same this time but, interestingly, it doesn't seem to be the case - it seems like the CE tree isn't getting updated at all, but Cooker is. Right now my advice would be to use the Cooker tree until official is released, then switch to either official or CE tree (assuming you don't want to carry on using Cooker).

devries: graphical install uses framebuffer access. in almost all cases if graphical install works than the vesa or vesafb driver will work for X, but it'll be slow.

what artee says will apply, though it doesn't help when you make a new directory... also, you could always use a lazy option, which is to go to the root of the partition and issue a: chown -R desireduser.desiredgroup * which will do the job. inelegant but effective. you could even script it. :)

console.perms is a pretty important file that sets a *lot* of permissions away from their default values to values which mdk's startup scripts expect. my guess is after the error, console.perms stopped being interpreted, so a bunch of permissions were set wrongly, and somehow some process could not access what it needed to (this is probably the cause of the /dev/ttyS0 error too) and this led to the rather misleading error message. It does seem a bit worrisome that this one file can cause *so* much trouble, though. I'd think there'd be a way to ensure at least a minimal boot with a non-functioning (or non-existent) console.perms...

I've been using xmame and gxmame from RPMs on 10.1CE for the last few days and they work great...what's the problem with the 10.0 RPMs?

thanks liquid, i was looking for that thread to help spunky out, couldn't find it. yep, 2.6.8.1-10 does manifest the problem. it'll happen in any kernel 2.6.8.1-mm2 or later (mdk's kernel fits the bill).

they trademark the name bluecurve. i believe several elements of it are available freely under the name 'wonderland'.

these rpms are built against Cooker. 10.1CE and Cooker are currently the same (assuming you update your CE from online repositories) so it will obviously work with both. The directory is called 2.7 because Goetz started building these several months back when it was still the 2.7 dev branch, and the directory serves as an urpmi repository so changing the name would force everyone to change their urpmi settings. The rpms will not work with 10.0 as they have several 10.1 specific dependencies.

devries: no, it doesn't mean that. since kernel-source doesn't require kernel, and kernel doesn't require kernel-source (this is correct), urpmi cannot know which kernel-source package to install. (besides, you can have any number of different kernels installed). Only options 4 and 5 in that list would work. option 1 is for kernel-multimedia-2.6, the "bleeding-edge" kernel in 10.1 (built by Svetoslav Slavtchev). option 2, kernel-tmb-source, is for the tmb kernel, the previous bleeding-edge kernel, which was built by Thomas Backlund. There's no real reason to use this over kernel-multimedia-2.6. option 3 is for the 2.4 version of Thomas's kernel, which there's not much reason to use any more either (unless 2.6 doesn't work for you and there's some reason tmb's 2.4 is better than stock mdk 2.4). only options 4 and 5 are appropriate for standard MDK kernels.

GNOME on my htpc box - spatial nautilus is a godsend with low resolution when all you really want to do is load videos. xfce on my poor old steam-powered laptop.

don't think so. they may use *different* ones from other projects...they seem to use ffmpeg by default, which I don't think transcode does.