fdisk-linux and its "m"... written for bsd ( written "better" of course...) something to replace sysinstall/bsdinstall/fdisk/gpart in the cases where they each return terse errors of some sort... due to GPT/MBR/GEOM/sysctl etc...
It is less intuitive than sysinstall, doesn't include an option to install to ZFS, and really as far as I am concerned is inferior in every way. Installing via command line is preferable... or using PC-BSD...
At first I didn't like bsdinstall but I've got used to it now. I got to the point where I found it better to install manually (especially when I got into ZFS as there really isn't another way) but now I just do the following - Go into bsdinstall to setup keymap + network if doing FTP install, drop to command line to set up disks/zpool/gmirror whatever, then back into bsdinstall to extract the OS.
I only use FreeBSD as a server OS so desktop stuff is way down my list of wishes (Not even on there in fact...). Believe it or not I'm quite happy with Windows 7 on my workstation
I would like to see the following:
High performance, production quality kernel iSCSI target (unlikely unfortunately)
Encryption commands in ZFS would be really nice. I probably would use encryption if I could just enable it on a dataset. The fact that I need to mess about with all the GELI stuff to encrypt my entire disks and then build ZFS on top of that puts me off.
Better disk drivers and integration between various GEOM layers so that disk errors aren't as likely to just hang IO, or to fix the fact that you can just pull a hot swap disk without ZFS even knowing about it until you manually offline. I believe zfsd may be in FreeBSD 10 to support auto-replace but until errors are handled properly all the way up the stack I can see a lot of things not getting handled correctly.
Native virtualisation stack (ala Linux's KVM/libvirt) would be awesome but I guess near impossible due to the amount of upfront and ongoing work required. At the moment the choice is really between either expensive commercial hypervisors or Linux servers.
I could probably think of a few other things that would be nice. It looks like booting from an exported ZFS pool (i.e. not already in /boot/zfs/zpool.cache) is being worked on which'll be nice. It's only really that that makes installing root-on-ZFS awkward.
Funnily enough it seems all mine are ZFS based. I've actually been using FreeBSD for over 10 years but for everything we use it for (web serving/email/DNS/radius/etc) it's been great. Hardware support could be better but we've never had any problems although we probably would if we shelled out for the latest servers. ZFS really is a major feature for FreeBSD but it's shown quite a few weaknesses. It'd be nice if FreeBSD 10 had a major new feature or improvement to shout about on release but I'm not aware of anything.
Native virtualisation stack (ala Linux's KVM/libvirt) would be awesome but I guess near impossible due to the amount of upfront and ongoing work required. At the moment the choice is really between either expensive commercial hypervisors or Linux servers.
If I was only able to change one thing, it would be to make more users realize that they have ownership of the OS. "FreeBSD: it's yours, too. If you see something that needs fixing, fix it, or help somebody else fix it."
If I was only able to change one thing, it would be to make more users realize that they have ownership of the OS. "FreeBSD: it's yours, too. If you see something that needs fixing, fix it, or help somebody else fix it."
I had a problem some days ago with snd_hda, and I had a great experience
of reading the code, debugging it and talking with mav@ ; the guy in charge of it.
He was nice and very patient with me, and together we made my system work, and
it resulted in a patch that will be merged soon.
That was fun!
So, I humbly thought that it would be so nice if dev guys like mav@ and others
write more prose, more often, to teach the rest of us how to do what they usually do.
There is /usr/ports/net-mgmt/wifimgr .
But it's not as full-featured as the the tools available in a Linux system for this purpose, especially on a laptop, where you might want a tray app to quickly identify the available networks and monitor the connection. I'd like to see something like wicd, which works well with xfce, kde, fluxbox, etc.
I'd really like to have new radeon GPUs supported. And since I'm already dreaming I'd also like OpenGL support. As stated before I am willing to chip in financially to get this done.
I'd really like to have new radeon GPUs supported. And since I'm already dreaming I'd also like OpenGL support. As stated before I am willing to chip in financially to get this done.
I think that FreeBSD is no longer as general purpose OS as it has been advertised so far. I'd like to see the focus shift more towards it's strengths which are IMO (including but no limited to) firewalls/routers, NAS, general purpose servers.
It has and always will be a research system.
This is its greatest strength. Those who wish to port it to newer CPUs and different architectures are people who understand what a system is capable of doing.
FreeBSD is for those that want efficiency and stability while truly pushing the limits of programming and hardware.
If you choose FreeBSD, you're stepping into a world that states, "You need to do it right, even if it takes time."
I'm going to agree that using 0 and 1 should stay as such.
One could relate 0 as open or even as the value of OR with 1 being closed and the value of NOR.
Now, it is of personal opinion and conjecture that starting software releases at $VALUE=0 is better than$VALUE=1.
With hardware, that should remain as is.
No reason why a bunch of nerds should put a number to waste.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
