iphitus

alias eth1 tg3

options eth1 disable_hw_scan=1

That's aliasing your ethernet card to eth1, not your wireless.

Further, you need more information than what you put in /etc/sysconfig/network-scripts/ifcfg-eth1 to get a network connection.

Compare wireless and wired....

Wired, 1) you plug it in 2) get an IP.

Wireless, 1) Associate with a wireless access point, 2) get an IP.

For both, you can't do 2 until you do 1. And like ethernet where you must plug in the cable, wireless needs some input before it can do #1, usually the network name, and a key if you have security.

Given that you're on fedora, you don't need the above ifcfg-eth1 file. Just have the driver loaded and use Fedora's networkmanager implementation. Networkmanager is patchy at best, but you're going to have a hard time with anything else on Fedora.

Good luck

James

Why do you want to do device driver programming?

Do you have a reasonable amount of experience in C?

Savage 1 was awesome, I can't wait to try Savage 2.

you'd have to split it, but it'd be ugly, and use up a lot of CD-Rs. There is no compression that can do what you want.

DVD burners are pretty cheap -- any attempt at splitting is going to take a lot of time and CD-Rs. If you've got a desktop, it might be better to just get a dvd burner.

install was actually impossible despite a bug having already been filed for earlier versions: https://qa.mandriva.com/show_bug.cgi?id=31356

Instead of chmodding the files to the new UID, you can just do:

usermod -u 1234 bob

To change bob's uid to his old one (1234).

Welcome to the board stephen!

this is going to be my first program i am going to install

Then you'll want to look at urpmi, I'm pretty sure there's a faq here in the faq section. It's a much simpler way of installing it.

urpmi make

no no no no! Wrong! Why do people still do this!

If a newbie is trying to compile an application on a brand new Mandriva install, you tell them to use urpmi. Not how to compile from the tarball. Nearly every time they are compiling because they dont know urpmi exists, or have not setup repositories. ettercap ought to have a package, otherwise there's also wireshark which should definitely have a package.

James

Another vote for GTKPod. You don't need to setup the iPod on a windows/mac computer to be able to use it, gtkpod can handle it just fine.

My iPod's never seen a windows box.

It really depends on the bug, and who it's assigned to, and to an extent, the project.

Filing a bug though, isnt a waste of time. At the least it provides a record of the issue. And in most cases, it can be followed up with a solution. Sometimes, it's not even a bug, but rather an error on the user's behalf, and things can be sorted out quickly.

One of the problems with Linux, is people tend to pick an itch, and scratch it. So various people will work on what they feel like. If they feel like coding tonight, thats what they'll do. If they feel like closing bugs, that'll happen. Some developers just don't like bug trackers, they'll avoid them as much as possible because they mean more work and at times, lengthy discussion to determine the problem. That's why intelligently researched bug reports tend to get fixed quicker, less work for the developer, whereas "X doesnt work!!!" will get ignored for a bit, or just receive "More info".

It depends on the project too. For Arch, the bug tracker is important, central, and a great deal of importance is given to dealing with the bugs there. The kernel though, generally runs off mailing lists. Try mailing to LKML and/or the relevant subsystem mailing list, you might get some more visibility.

But I know what it's like. I posted a bug to NetworkManager well over 6 months ago and it has yet to be so much as triaged. Or another dead simple one which i posted to Mandriva which has been open for a similarly long amount of time.

And I'm guilty too, I've got some bugs assigned to me which are a tad crusty, simply because I don't deem them important, or I'm not motivated to deal with them. Though I did make an effort to get rid of a lot of them the other week.

So then Firefox would be configured to use squid as a proxy server?

But then couldn't each user then just edit the Firefox preferences to connect directly to the internet without going through squid, and could then see any sites they wanted??

You could adjust the system's networking to disallow direct internet access.

From what I have seen and read, I have this personal fear that KDE4 is going to turn out to be yet another piece of worthless eye candy just like compiz et al.

Guess you have not looked at KDE4 at all. Have you seen the work into the underlying foundations? KDE4 is an impressive piece of work that will allow some amazing sofware. Time will prove this as people learn to exploit the capabilities of KDE4. It's not just some desktop environment, it's a whole foundation of libraries to develop on. Phonon, Solid, ThreadWeaver, Decibel, etc. At first sight, they're impressive.

As I sit at my desk here I am surrounded by 5 hardware items that don't work, either properly, or at all, with Linux (any version) and I am using a laptop that will either shut down 1 time out of 6 (Mandriva) or will not shut down at all (any other distro) and yet I have this awful feeling that vast amounts of human resources are going to be ploughed into plasmoids (whatever the hell they are) and the rest, whilst such utter basics as I have just described, have gone, or will go, unaddressed for ever.

Please tell me what these hardware items are. I'm _really_ curious. I betcha at least some of them will work.

If Linux is to gain the credibility it deserves then it needs people to continue to improve on the basics, not to go off on wild and useless flights of fancy in some daft effort to emulate Windows (that is all that compiz was anyhow).

There's no central body. People work on what they want to work on. Nobody can marshall people and say "stop working on compiz, the kernel needs work". Nobody can redirect or move those human resources, they're working of their own will.

mutt + msmtp.

msmtp is very easy, just take an example config and modify it as needed.

James

modules.conf is old, the documentation you refer to above is dated. Ignore it.

Got nothing to do with cpufreq. The mandriva default laptop_mode configuration must set the drive spindown time. Poke around at the .conf's in /etc/laptop-mode (assuming mandriva puts it in the same place as arch).

Also, check if kpowersave controls hdd spindown time.

James

Does mandriva use laptop_mode or any form of hdd power management/sleep?

This exact thing happens to me when I set my drive to spin down after X minutes of inactivity. I've got pretty aggressive power saving settings as my battery is old and I get a pause when the hdd spins up again. Whatever program was trying to read from it is effectively put on hold while the drive spins up.

If it's a multi distro bug, file a bug on the kernel bugzilla. Mandriva will probably just ignore it and wait for an upstream fix.

If the bugs on either havnt been triaged/commented on after a few weeks, bug them and provide alll the information they ask for. If they still dont fix it, keep bugging them.

Not necessarily. I always used to update kernels, and now I have a bug report open, because every kernel since 2.6.12.x means my system doesn't shut down properly.

Where's that bug? Nag at it! 2.6.12 something like 2 years old!

If it's at Mandriva, ditch it and repost it at the kernel bugzilla. Repeatedly bug them until it's fixed.

Epiphany with webkit is pretty awesome too. If they've completed their python support, then I'm going to have a lot of fun churning out some great extensions.

Havn't done much for a while.

Best place to start is as linked above, kernelnewbies.org

Robert Love's Linux Kernel Development is a must read if you're fairly new to the kernel.

Otherwise, find and itch and scrach it. If there's something you'd like to fix/improve, take a look at that, read the documentation, otherwise there's lots on kernelnewbies.

Don't waste time on kernel patchsets. It's not really kernel development, you don't learn a great deal about the kernel, and you never actually get much coding done.

James

Don't know what dated version Mandriva distribute, but now Compiz Fusion = Compiz.

Previously there was one, Compiz, which was then forked to make two, Compiz and Beryl, which later merged back as Compiz Fusion.

kde/gtk-window-decorator are the responsibility of the Compiz Fusion project, hence KDE washing their hands of it. First, try the latest version, it may be an old bug. If it remains an issue, gather what information you can and file a bug with the compiz fusion project or Mandriva.

James

It's probably because it's for an older version of Gnome, and hence causing the crashing. I had this as well. I just made sure I was not using a theme which was for an older version of Gnome.

GTK != Gnome

GTK Themes != Gnome.

But yeah, poorly made themes do crash applications, and it's possible that it may be a gtk theme engine that openoffice doesnt co-operate with so well.

James

Could it be a limitation of the filesystem used on the CD?

Then just split the file in half, and then burn it all to a dual layer disc.

4.6gb is also very big. Unless it's something like 4 hours long, then I'm sure you can transcode it and make it substantially smaller. How long is the recording?

This thread is out of date, and doesnt have any reason to stay.

Anyone got any objectives if I unpin and let it fade into history?

The term getting your feet wet hasn't any meaning here!!

 

After patching and compiling your kernel which, unless you really need specific changes, can be an old compile. After being installed the kernel is stable and gives you a good stable start up splash and shut down splash. You can even put in a nice grub jpeg at boot if you like without too much trouble.

And splashy can do it without any kernel modifications. That in itself is a great advantage. It's not a matter of "getting your feet wet" -- it's a matter of practicality.

Bootsplash is a _dated_ and now inferior implementation. Not to take anything away from it - it served reliably during it's time, but now it's time has passed.

Mandriva (and it's derivatives) is the last distro I know of that uses/has used bootsplash (I havn't checked the latest releases).

Requiring a kernel patch is a great disadvantage, and in bootsplash' current implementation, these include:

• It embeds a jpeg loader in the kernel. This is a dirty hack by today's standards, where there are multiple implementations that prove this is not neccesary
  
• Putting a jpeg loader into the kernel raises extra security concerns.
  
• In kernel space, there is _no_ memory protection. A bug in the jpeg loader, could potentially corrupt ram, cause data loss, difficult to trace bugs elsewhere, or simply crashes. Another reason why it should be in userspace
  
• Because of the above, it will never be merged into the vanilla kernel
  
• The bootsplash implementation is poorly documented, and as the original authors have long ditched it, there exists nobody with the knowledge _and_ enthusiasm to maintain it. The fbsplash author attempted to work it out, and decided it would be easier to write his own (fbsplash).
  
• In the kernel it is feature limited, difficult to configure, requires very invasive modification to work on any distro
  
• It can, and has been done better, with other implementations putting only the bare minimum in the kernel (fbsplash+fbcondecor), and others nothing (fbsplash/splashy).
  

So in short. It sucks by _today's_ standards.

Where do you think splashy came from? :mellow:

People who realised that bootsplash has big flaws, and a new implementation was needed.

It seems to me that folk only want to have a dig without really understanding anything at all behind bootsplash

Not making a blind dig at all. I know bootsplash _very_ well, along with it's successors. From maintaining patches to updating initscripts. I _could_ keep it up to date, but it's not worth it.

Put your time towards another implementation such as splashy or fbsplash/fbcondecor. You could pursue bootsplash, it's open source after all, but expect a lot more of the above.

James

bootsplash is a dirty ugly hack. It embeds a custom and poorly documented jpeg loader into the kernel. It has no hope of ever being merged.

bootsplash was always primitive anyway, never had a great deal of capability and was always limited by the fact it depends on a kernel patch.

Implemented correctly, splashy is far superior, and far cleaner. It's more capable, and importantly, its entirely in userspace. fbsplash isnt too shabby either.