just brew it! wrote:I guess I'll have to figure out how to disable whatever new voodoo it is they've implemented, so that I can just configure things manually in /etc/network/interfaces (the old fashioned way).
bitvector wrote:just brew it! wrote:I guess I'll have to figure out how to disable whatever new voodoo it is they've implemented, so that I can just configure things manually in /etc/network/interfaces (the old fashioned way).
The Debian way.
axeman wrote:8.10 feels a little bit like using Debian Testing - okay if you're relatively savvy, but not really ready for prime-time.
bitvector wrote:Whoa, whoa, whoa. As a seven year user of debian testing, I'd demand you take that back. The magnitude of radical half-baked changes imposed in Ubuntu for the sake of novelty and "cutting edge features" is way way greater than Debian testing. The further Ubuntu diverges from Debian upstream, the more they're getting to be like Fedora -- always shipping broken stuff because they're trying to move just a little too fast into new functionality.
axeman wrote:I would say pushing gcc 4.3 into testing causing me to be constantly pointing the gcc symlink back to 4.1 everytime I update so I could compile kernel modules again is pretty half baked. But then again, maybe I'm just being picky and should have just ignored it for the past 8 months.
axeman wrote:I haven't found Ubuntu to be nearly as broken as Fedora.
Sure they are a little bleeding edge, but they have a big enough community to keep things (mostly) working. Then again, I'm still having a headache with Samba on every other 8.04 machine I've touched and I've NEVER had that problem with Debian. At least 8.10 fixed that for me.
axeman wrote:bitvector wrote:What modules were you compiling manually? If you were using module-assistant, a Debian approved kernel module compilation mechanism, it would have handled all of that properly and automatically.
axeman wrote:WRT vmware on Debian, you should definitely check out vmware-package. A friend of mine is a maintainer for it and it really makes it much more Debian-style in the layout, packaging and module compilation. It takes the input vmware tarballs (for player, server, workstation, etc.) and makes modular .deb packages out of them. It can take the any-any updates or kernel modules from the releases and make module-assistant compatible source packages. I've been using it for a long time and it makes it all quite painless. I just do m-a a-i vmware-any-any-kernel-source and, boom, the modules are compiled, bundled into a module deb, installed, and good to go. The only snags I've run into is vmware's fault: periods where the any-any updates and vmware support aren't actually keeping up with newer kernel compatibility.VMWare.
axeman wrote:I agree with that. I've been using ntfs-3g from Debian experimental, though, and find it's pretty recent. The current one is about a month behind (using the Sept. 1.2918 release rather than the release 1.5012 of a couple weeks ago) and gets pretty quick updates.Also compiling recent versions of ntfs-3g, since most distros are woefully behind on updating this, including Ubuntu.
henry44 wrote:I tried 8.10 (both 32 and 64 bit versions) on a Lenovo x200 and had to go get video, NIC and wireless drivers. The sound I could never get working. Since this laptop is basically my PDA, cell phone and mp3 player, the sound issue is a deal breaker.
srg86 wrote:Probably assumptions about exactly which instructions are supported by non-mainstream processors and Debian Etch is getting it wrong whilst Ubuntu 8.10 is either actually reading the cpuflags correctly or guessing it right. It always used to be a problem with the Via EPIA boards so I can easily believe a Cyrix 6x86 having problems.Also, one thing I don't get, Ubuntu boots fine (but slowly of course) on the Cyrix 6x86 but Debian Etch hangs as soon as the kernel starts.
notfred wrote:Probably assumptions about exactly which instructions are supported by non-mainstream processors and Debian Etch is getting it wrong whilst Ubuntu 8.10 is either actually reading the cpuflags correctly or guessing it right. It always used to be a problem with the Via EPIA boards so I can easily believe a Cyrix 6x86 having problems.