XFCE 4.8 is Linux only ?

pelmen

Member

Reaction score: 9
Messages: 39

Hi,

from today announce at xfce.org:
http://www.xfce.org/about/news/?post=1295136000

We hope that everyone will enjoy this release as much as we do. Sadly, this will not be the case as the folks using any of the BSD systems will notice a sudden loss of features. We think that this announcement is a good opportunity to express our disagreement with the recent "Linux-only" developments in the open source ecosystem, especially with regards to the utilities we need in desktop environments.
What is mean, this is end of XFCE on BSD ?
Ive use it on Asus EEEPC and FreeBSD ;(
 

SirDice

Administrator
Staff member
Administrator
Moderator

Reaction score: 8,021
Messages: 31,497

It means not all functionality will work. That's all.

It's not clear what this missing functionality is though. At least I couldn't find anything in the changelog.
 

jfrelinger

New Member


Messages: 8

Most likely this is about moving from hal to udev. Xfce has removed using hal and moved to using udev for managing removable devices. Unfortunately udev is linux only. So plug-able devices will no longer work.
 

tingo

Daemon

Reaction score: 430
Messages: 2,105

jfrelinger said:
Most likely this is about moving from hal to udev. Xfce has removed using hal and moved to using udev for managing removable devices. Unfortunately udev is linux only. So plug-able devices will no longer work.
(shrug)
I have always been running without HAL, so no change for me. I am perfectly capable of mounting removable devices myself, thankyouverymuch. :e
 

phoenix

Administrator
Staff member
Administrator
Moderator

Reaction score: 1,260
Messages: 4,099

Now, if we could only get the major desktop developers to look at our wonderful devd.conf(5) system, which has been around far longer than udev, hal, devfs, dev-thingy-of-the-month, and provides all the features needed to make this work.

Of course, they've had almost 10 years to work on this ...
 

sossego

Retired from the forums

Reaction score: 144
Messages: 1,557

I think that vermaden- not sure right now- has a post on daemonforums for using X without hal or dbus.

It's a shame that the different BSD projects don't stop and talk to each other.
Id est: Ask the OpenBSD Xenocara project how they are working around the udev problem.
 

Crivens

Moderator
Staff member
Moderator

Reaction score: 986
Messages: 1,986

phoenix said:
...
Of course, they've had almost 10 years to work on this ...
My sarcasm-o-meter just gave me a nice *ping* here.

This frogger-type development must be an evil plan from some sinister company we all know, right? That's what I suspect from time to time.

But since I also tend to exorcise hal and other nanny-software from my installations, I will also shrug this off. Seriously, who will want software which depends on polling the CD drive for media? Wasn't the DiskIsChanged(yes,no,maybe) from DOS enough?

But I am glad that I am not alone with my dislike for any "that-os-only" approach.
 

phoenix

Administrator
Staff member
Administrator
Moderator

Reaction score: 1,260
Messages: 4,099

devd is not a polling interface. It is an event-driven hardware framework. It is what the Linux guys have been trying to figure out for almost 10 years now, but still haven't.

When you insert a piece of hardware, the hardware sends a notification to the kernel via devd. devd then fires off whatever action you configure it to. No polling involved.

Pretty much anything you can do with HAL/udev on Linux, we can do on FreeBSD ... but done properly.

What sucks is that all the major desktop-oriented software devs are Linux users, so they don't know how things should work. :) So they come up with these kludges like udev, udisk, upower, uwhatever.

Read the devd.conf(5) man page. You'll be surprised what you can do with it.
 

Crivens

Moderator
Staff member
Moderator

Reaction score: 986
Messages: 1,986

Sorry, maybe I should have been more clear. You are knocking on open doors here.

That udev is a step in the right direction is correct and it should be applauded.
When I was ranting about polling, that was aimed at hal. That it can not be compared with devd is also clear, but I am still a little bit spoiled by the first OS I used.
 

SirDice

Administrator
Staff member
Administrator
Moderator

Reaction score: 8,021
Messages: 31,497

Crivens said:
That udev is a step in the right direction is correct and it should be applauded.
Actually it isn't. Stepping away from HAL is the right direction, so far so good. But moving to udev doesn't help us (FreeBSD users) at all. That would mean we'll end up with no support for things like auto mounting USB sticks. At least not from a user's point of view.

HAL may not have been the best solution but at least it worked on both Linux and FreeBSD.
 

pkubaj

Aspiring Daemon

Reaction score: 101
Messages: 555

Can't FreeBSD devs port udev just like KMS and GEM are getting ported from Linux kernel? I've heard they're to hit 9.0-RELEASE.
 

graudeejs

Son of Beastie

Reaction score: 695
Messages: 4,615

pkubaj said:
Can't FreeBSD devs port udev just like KMS and GEM are getting ported from Linux kernel? I've heard they're to hit 9.0-RELEASE.
Even if they could, the question is how long udev will be around....
You know, linux is not about stability
 

Crivens

Moderator
Staff member
Moderator

Reaction score: 986
Messages: 1,986

Correct, I use the term "frogger-style-development" for such attempts.
You jump around from one solution to the next and end up under a carwreck...

Maybe a layer to translate devd events to other types (udev, even hal) would be a better hack to get that software currently produced for linux (from today, maybe next week) running. And I mean hack, but it would be contained in one place. A translator from hal/udev to devd, and licenced GPL, would even be better. Some app developers might go for it, thus forcing devd into linux :e

Udev may be a step in the right direction, but when standing under a falling piano, each direction is a step in the right direction (my opinion regarding hal). That step may bring you over the open manhole, but that did not stop willy coyote, did it?
 

pkubaj

Aspiring Daemon

Reaction score: 101
Messages: 555

killasmurf86 said:
Even if they could, the question is how long udev will be around....
You know, linux is not about stability
Ok, but nobody's going to support something that is FreeBSD-only. Linux is now considered mainstream OS, along with Mac OS X and Windows. Furthermore it's found in plenty of modern PDAs and cellphones. FreeBSD is considered a very niche OS with extremely small user base. It doesn't really matter for server installations, but in order to have a usable desktop, we need things like USB automounting (at least I do) and hardware drivers (and new Intel GPUs don't work right because of the lack os some Linux kernel features). You may like or not (I don't) but in order to make FreeBSD survive on a desktop, the developers MUST port these features from Linux, or patch existing software to work with FreeBSD kernel features (like devd, of which I didn't even know). Otherwise, I'm afraid, many users will leave FreeBSD (like me). I really love FreeBSD and can cope with some difficulties, but if FreeBSD really stays behind, then I'll have no choice. And I think it's already happening. E.g., I really would like to install FreeBSD on my notebook, but I can't, since it features Intel GMA HD GPU, for which there are no FreeBSD drivers, because there's no KMS support. And, sadly, I'm writing this from Linux.
 

phoenix

Administrator
Staff member
Administrator
Moderator

Reaction score: 1,260
Messages: 4,099

pkubaj said:
Can't FreeBSD devs port udev just like KMS and GEM are getting ported from Linux kernel? I've heard they're to hit 9.0-RELEASE.
The KDE devs have the right idea; unfortunately, there hasn't been a lot of follow-through outside of Linux yet. Phonon and Solid is how it should be done (interface layers that take pluggable backends; so no matter how often the backend changes, higher level software never changes).

We (FreeBSD KDE users) don't need a port of udev, or a port of HAL, or a port of HAL-like-thingy-of-the-month. What we need is for someone to write up a proper devd-based backend for Solid (or a similar compat layer for non-KDE users). What's funny, is that the KDE devs have written udev, HAL, upower, u-whatever, etc backends multiple times for Solid, yet no one has written the one devd-backend we really need.

devd, powerd, and sysctl together have been doing what devfs, new-devfs, udev, HAL, udisk, upower, etc were designed to do ... but don't quite do yet.

It's too bad we don't have any heavy KDE users, who've become programmers, who use FreeBSD, who could whip this up. Would be interesting to compare how long it takes to make a devd backend for Solid, compared to how long it took each of the rewrites of the udev/hal/whatever backends. :)
 

phoenix

Administrator
Staff member
Administrator
Moderator

Reaction score: 1,260
Messages: 4,099

pkubaj said:
Ok, but nobody's going to support something that is FreeBSD-only.
Yet, you expect the world to support something that is Linux-only? It's that idea right there that is the problem.
 

pkubaj

Aspiring Daemon

Reaction score: 101
Messages: 555

phoenix said:
Yet, you expect the world to support something that is Linux-only? It's that idea right there that is the problem.
As I wrote, Linux is considered a mainstream OS. Besides FreeBSD has about 0.01% market share, while Linux has 1%, which is 100 times more. On top of that, Linux is found in a lot of embedded devices (for example my Nokia N900 has Maemo - Debian-based distro, there are also lots of Android-based phones and HP is also trying hard with webOS), but it's not the case with FreeBSD. Linux is WAY more popular and either FreeBSD adjusts or it will disappear from desktop / laptop computers. Don't get me wrong, I'm not pro-Linux, but this is reality. I think that the real all-time solution would be some Linux kernel calls compatibility layer. I don't think it would be that difficult to do, since we already have software compatibility layer, and both OSes are Unix / Unix-like, but I know nothing about programming, so I may be wrong.
 

ckt1g3r

New Member


Messages: 11

killasmurf86 said:
Even if they could, the question is how long udev will be around....
You know, linux is not about stability
If you say that Linux is not stable it probably never experienced slackware, debian, centos (redhat). I use Slackware Linux and FreeBSD, and i not found any differences in stability.

Sorry for my bad english.
 

UNIXgod

Daemon

Reaction score: 200
Messages: 1,088

pkubaj said:
As I wrote, Linux is considered a mainstream OS. Besides FreeBSD has about 0.01% market share, while Linux has 1%, which is 100 times more. On top of that, Linux is found in a lot of embedded devices (for example my Nokia N900 has Maemo - Debian-based distro, there are also lots of Android-based phones and HP is also trying hard with webOS), but it's not the case with FreeBSD. Linux is WAY more popular and either FreeBSD adjusts or it will disappear from desktop / laptop computers. Don't get me wrong, I'm not pro-Linux, but this is reality. I think that the real all-time solution would be some Linux kernel calls compatibility layer. I don't think it would be that difficult to do, since we already have software compatibility layer, and both OSes are Unix / Unix-like, but I know nothing about programming, so I may be wrong.
http://comments.gmane.org/gmane.os.freebsd.stable/73977
^^This.......

So who in their right mind would ever want to create software that is not portable? Maybe someone who was taught incorrectly.
 

DutchDaemon

Administrator
Staff member
Administrator
Moderator
Developer

Reaction score: 2,840
Messages: 11,317

ckt1g3r said:
If you say that Linux is not stable it probably never experienced slackware, debian, centos (redhat). I use Slackware Linux and FreeBSD, and i not found any differences in stability.
That's not the type of stability we're talking about. This is not about crashing computers or something like that. It's about feature / architecture / back-end stability (or 'un-changeability' if you like), a stable code base -- stability that programmers can rely on when designing applications and considering back-ends for them. So not following every latest fad and fashion, but choosing a solid base to build on.
 

Crivens

Moderator
Staff member
Moderator

Reaction score: 986
Messages: 1,986

phoenix said:
The KDE devs have the right idea; unfortunately, there hasn't been a lot of follow-through outside of Linux yet. Phonon and Solid is how it should be done (interface layers that take pluggable backends; so no matter how often the backend changes, higher level software never changes).
I was not aware of that part in KDE. Well, I could give it a try, provided the footprint is not too big. For now, XFCE has the right balance when features and startuptime / memory / disk space is considered. Oh, and the close button goes top-left, that's a must. ;)

*ducks for cover as the flamewar about window managers/DEs is prepared*
 

DutchDaemon

Administrator
Staff member
Administrator
Moderator
Developer

Reaction score: 2,840
Messages: 11,317

Crivens said:
*ducks for cover as the flamewar about window managers/DEs is prepared*
Oh, I have a close button too ...
 
Top