Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
X 7.1 still testing on x86/amd64 because of BINARY drivers?
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 14, 15, 16  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  

Should X 7.1 still be testing only because of closed source drivers (ati/nvidia)?
Yes!
55%
 55%  [ 254 ]
No!
37%
 37%  [ 172 ]
I don't care!
6%
 6%  [ 28 ]
Total Votes : 454

Author Message
madman2003
Apprentice
Apprentice


Joined: 20 Feb 2005
Posts: 178

PostPosted: Tue Aug 29, 2006 12:24 pm    Post subject: Reply with quote

Matteo Azzali wrote:
madman2003 wrote:

In my case, a while was a little over a month. Xorg 7.0 had some problems for me, 7.1 didn't. My reasoning is that if all users are confronted with the downsides of non-free drivers, then they will at least be aware of it, instead of coverering it up. It might also give nvidia/ati a reason to speed things up a bit, because there is actually no good reason to delay a minor change for months, besides having no reason to do it (no distro was really using it and that's because the drivers aren't there, vicious circle).


You needed xorg-7.1, ok. Do you know the /etc/portage/package.keyword mechanism? Is so difficult to put in there the
packages you needed? Or do you think that 3D users in ~ branch didn't used package.mask to prevent xorg-7.1 from
compiling on their system? Whenever there are large groups of peoples with different needings there are developers
choiches to try to make much possible people happy, and is still 100% sure someone will complain. Sad.
ABI change isn't a minor change, at least for Nvidia they specified that just recompiling the drivers wasn't enough
(probably there's some assembly involved,and in this case they had to modify the asm sources for every supported platform
and for legacy drivers).And you can be sure that before any drivers release there's some testing as they don't want
that users boards implode for a distraction error.....

Also in the current situation (where directX 10 graphics archs -r600 and G80 -are round the corner, and isn't 100% sure
those boards will be easily compatible with OpenGL) I would avoid to push/force manufacturers to do anything, in the hope
they'll continue to support linux community, as we still haven't viable alternatives for any of the boards in the shops.

In the end I would like just to remember that ATI x1000 boards owners waited more than 6 months to have drivers,
and there wasn't all this fuzz about.


You seem to misunderstood me, recent events looks like the progress of gentoo does depend on non-free software, which is troubling. I would have made the choice the other way around, let the default be 7.1 in stable branch and let people mask it if they need the non-free drivers.

I guess it depends on how the social contract is interpreted, but i've had my say.

So don't think me as a troll or anything like that.
Back to top
View user's profile Send private message
Matteo Azzali
Retired Dev
Retired Dev


Joined: 23 Sep 2004
Posts: 1133

PostPosted: Tue Aug 29, 2006 12:45 pm    Post subject: Reply with quote

I understood perfectly, but my choiche would have been different: let xorg-7.1 be completely masked (even to unstable
branches) and let people really needing it to unmask it by hands, cause less than 10% of the users had issues with xorg-7.0
(and is a gentoo policy to avoid any "ng" package with issues that's not really needed).

So I feel that actual situation is something to try to satisfy as much people as possible (not that I'm unhappy with it, I just
masked xorg-7.1 in my package.mask till 3 days ago).
_________________
Every day a new distro comes to birth. Every day a distro "eats" another.
If you're born distro, no matter what, start to run.
---- http://www.linuxprinting.org/ ---- http://tuxmobil.org/
Back to top
View user's profile Send private message
Kensai
Guru
Guru


Joined: 09 Mar 2004
Posts: 568
Location: Puerto Rico

PostPosted: Tue Aug 29, 2006 5:58 pm    Post subject: Reply with quote

R!tman wrote:
I'm pretty sure you're wrong concerning at least Debian. As Debian doesn't even include any binary stuff, including nvidia and ati drivers, I very much doubt the devs would block anything because of things they don't even support ;-).

Go and see that they are now starting to build xorg 7.1 after the nvidia driver has been upgraded, man almost every distro there is didn't upgraded because of nvidia, ati.
_________________
Gentoo: Gigabyte: nFORCE 2: nVIDIA GeForce 6600: AMD Athlon XP 3200+
Perspective of a Thinking Human Being
Back to top
View user's profile Send private message
Slippery Jim
Apprentice
Apprentice


Joined: 08 Jan 2005
Posts: 264

PostPosted: Wed Sep 06, 2006 6:07 pm    Post subject: Reply with quote

Matteo Azzali wrote:
I understood perfectly, but my choiche would have been different: let xorg-7.1 be completely masked (even to unstable
branches) and let people really needing it to unmask it by hands, cause less than 10% of the users had issues with xorg-7.0
(and is a gentoo policy to avoid any "ng" package with issues that's not really needed).

So I feel that actual situation is something to try to satisfy as much people as possible (not that I'm unhappy with it, I just
masked xorg-7.1 in my package.mask till 3 days ago).


I agree that if 7.1 is to be masked from stable because of the binary drivers, then it should not be masked by ~arch. Masking incompatible xorg versions for the binary drivers is a policy decision. ~arch is for testing, not for policy masking. If 7.1 is stable, then it should be in the stable tree, whether it is masked by some other mechanism or not.

That being said, I don't think that newer xorg versions should be hard masked because of the binary OpenGL drivers. If the user wants to use a binary driver, then we can code the dependency for a compatible xorg version into that binary driver's ebuild, and let portage sort it out. That way, both camps can have what they want.

Also, consider that if you hard mask binary incompatible xorg versions from stable, then you can't just mask 7.1; you have to mask all versions greater than 7.1 as well. What if ATI and nVidia never release a binary driver for 7.1? Should 7.0 be the last stable xorg in portage?
Back to top
View user's profile Send private message
vipernicus
Veteran
Veteran


Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.

PostPosted: Wed Sep 06, 2006 6:20 pm    Post subject: Reply with quote

Why does Gentoo have to be bleeding edge and update immediately?
_________________
Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog
Back to top
View user's profile Send private message
VinzC
Watchman
Watchman


Joined: 17 Apr 2004
Posts: 5098
Location: Dark side of the mood

PostPosted: Wed Sep 06, 2006 7:26 pm    Post subject: Reply with quote

Kensai wrote:
Go and see that they are now starting to build xorg 7.1 after the nvidia driver has been upgraded, man almost every distro there is didn't upgraded because of nvidia, ati.

The fact they waited doesn't necessarily mean they're going to put binary stuff into their distribution. Maybe they waited simply to avoid running into the problem we had with Gentoo, after all? Maybe they waited long enough because they know that even if they don't include binary drivers there are people who manage to install ATI/nVidia drivers in Debian? And if there are people who do that maybe they just reasonnably wanted to take care of them anyway? Don't you think it could be a reasonnable explanation?
_________________
Gentoo addict: tomorrow I quit, I promise!... Just one more emerge...
1739!
Back to top
View user's profile Send private message
Slippery Jim
Apprentice
Apprentice


Joined: 08 Jan 2005
Posts: 264

PostPosted: Wed Sep 06, 2006 7:51 pm    Post subject: Reply with quote

vipernicus wrote:
Why does Gentoo have to be bleeding edge and update immediately?


No one's saying it has to be. The question, as I understand it, is whether an otherwise stable xorg should be masked because of the binary driver dependency.
Back to top
View user's profile Send private message
Kensai
Guru
Guru


Joined: 09 Mar 2004
Posts: 568
Location: Puerto Rico

PostPosted: Wed Sep 06, 2006 9:56 pm    Post subject: Reply with quote

VinzC wrote:
Maybe they waited long enough because they know that even if they don't include binary drivers there are people who manage to install ATI/nVidia drivers in Debian? And if there are people who do that maybe they just reasonnably wanted to take care of them anyway? Don't you think it could be a reasonnable explanation?

Thank you very much, that proves my point correct. :wink:
_________________
Gentoo: Gigabyte: nFORCE 2: nVIDIA GeForce 6600: AMD Athlon XP 3200+
Perspective of a Thinking Human Being
Back to top
View user's profile Send private message
boroshan
l33t
l33t


Joined: 16 Apr 2003
Posts: 730
Location: upside down

PostPosted: Fri Sep 15, 2006 10:25 am    Post subject: Reply with quote

Slippery Jim wrote:
The question, as I understand it, is whether an otherwise stable xorg should be masked because of the binary driver dependency.


Not whether. Whether assumes it's an either/or question, and it's not. It's a trade off. Mask for too long and Gentoo falls behind. Release too early and you get mass breakage and flaming because an understanding of the issues hasn't had time to propagate.

Framing the issue as an either/or decision is a false dichotmy. It's a useful one for the hardline FSF faction who can frame the debate along the lines of
Quote:
D00d! NVidia and ATI might NEVER release another binary. ALL progress on Gentoo will stop FOREVER and it will all be YOUR FAULT!!!

It's effective as an appeal to the emotions, but it's nevertheless misleading and fallacious.

If we're sensible the questions to ask are "what sort of masking period maximises works best". Or possible "how do we know when it's time to unmask a package?" Or maybe "what sort of timescale is it reasonable to allocate for proprietory drivers?"
_________________
Don't let THEM immanentize the Eschaton!
Back to top
View user's profile Send private message
batistuta
Veteran
Veteran


Joined: 29 Jul 2005
Posts: 1384
Location: Aachen

PostPosted: Fri Sep 15, 2006 11:12 am    Post subject: Reply with quote

boroshan, I kind of tend to agree that "whether" is not *that* innapropriate. You could move it to testing it from day one without a lot of serious problems. You will get a blocked dependency reported by portage and your update will break. You then go, mask those packages (or copy-paste the mask section kindly posted by a Dev *in advance*) and in 5 min your are ready to go. FSF fanatics are happy. I will be unhappy for 5 min, but I take it. Gentoo doesn't "fall behind", only users that want to (like me and you). It's a nice compromise (then I agree with your view, it's a compromise).

Now the problem was that this time it was handled by a gorilla with a gun. The update process went half way, breaking in the middle, leaving hundreds of people with a broken system. You reboot (I know linux doesn't need to reboot, but I still pay for my energy bill), and then your system doesn't start X. So have fun searching for a workaround on the net without a graphical browser. So you fire up text-lynx from the command line (unless you have framebuffer setup) and copy by hand all text you find, trying to filter the bitching from furstrating users.

My point: it's impossible to make everyone happy. So if you are a Gentoo Dev, do whatever you think is the best. Someone wants to make a FSF speech? Fine! But block packages accordingly, you won't make your point any stronger by breaking 10000 peoples' box, you will only end up with a 16-page long thread.
And I'm sure the Devs *didn't* want to break our boxes. Or really believe so.
So instead of saying all this -what I believe- unrelated and nothing to do arguments about free software and nvidia and binary and so on, some should admit "this was done wrong, boxes where broken unintentionally". Because
- There was a way to handle this without breaking Gentoo boxes
- I really can't believe that any Dev intentionally did this if there were a workaround. If I found that someone did it intentionally, this dev would lose all my respect. And I respect very much all Gentoo volunteers for their hard work.

I'm sure a lot of Gentoo users don't care about using closed source drivers, they care about someone breaking their computers. And so do I. Next time, it will be done better.
Back to top
View user's profile Send private message
boroshan
l33t
l33t


Joined: 16 Apr 2003
Posts: 730
Location: upside down

PostPosted: Fri Sep 15, 2006 2:14 pm    Post subject: Reply with quote

Quote:
And I'm sure the Devs *didn't* want to break our boxes. Or really believe so.

Agreed.

Quote:
So instead of saying all this -what I believe- unrelated and nothing to do arguments about free software and nvidia and binary and so on, some should admit "this was done wrong, boxes where broken unintentionally".

Agreed again. This would have saved a lot of angst and flameage. I think it might still.

The trouble is, we get the FSF lunatic fringe chanting ha ha! suxxors! serves you ri-i-ght. and the devs jump in to tell us about free software and social contracts, but I didn't see any of them setting the trolls straight. After a while, that makes it look as if there is implicit endorsement of the loony viewpoint by the developer community. And that's when people start getting angry.

A simpled "whoops, sorry" would have saved a lot of trouble.

Quote:
I'm sure a lot of Gentoo users don't care about using closed source drivers, they care about someone breaking their computers. And so do I. Next time, it will be done better.

I'm sure it will. And if it isn't, this discussion can always be re-opened.
_________________
Don't let THEM immanentize the Eschaton!
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 54028
Location: 56N 3W

PostPosted: Fri Sep 15, 2006 5:22 pm    Post subject: Reply with quote

boroshan,

Rule 1. Do not feed the trolls
Rule 2. There is no rule 2 - see rule 1.

There was a lot of correspondence on the -dev mailing list about this topic. I think the archive is public, so you can see it all for yourself. The option to keep Xorg at a version that worked with binary drivers was only an issue for x86 and amd64. The are no binaries for other arches.

It was more a 'least worst' choice than a 'best choice' and it was made without the benefit of the hindsight we have now. Will the same decision be made next time this happens is the question to be asking now?

What are the lessons learned?
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Ibn al-Hazardous
Tux's lil' helper
Tux's lil' helper


Joined: 02 Sep 2004
Posts: 133
Location: Somewhere deep in the desert.

PostPosted: Fri Sep 15, 2006 5:53 pm    Post subject: Reply with quote

NeddySeagoon wrote:
boroshan,
What are the lessons learned?


That mutual blocks are good? ;)
_________________
/Ibn
Back to top
View user's profile Send private message
Slippery Jim
Apprentice
Apprentice


Joined: 08 Jan 2005
Posts: 264

PostPosted: Sat Sep 16, 2006 1:57 am    Post subject: Reply with quote

NeddySeagoon wrote:
boroshan,

Rule 1. Do not feed the trolls
Rule 2. There is no rule 2 - see rule 1.

There was a lot of correspondence on the -dev mailing list about this topic. I think the archive is public, so you can see it all for yourself. The option to keep Xorg at a version that worked with binary drivers was only an issue for x86 and amd64. The are no binaries for other arches.

It was more a 'least worst' choice than a 'best choice' and it was made without the benefit of the hindsight we have now. Will the same decision be made next time this happens is the question to be asking now?

What are the lessons learned?


Ouch!

I wasn't trying to be a troll. Sorry if it sounded like I was.

I was trying to paraphrase the poll question, because it sounded to me like vipernicus read it differently than I did.

The poll asks if 7.1 should remain in testing because of binary drivers, which I read as "should it remain not stable?".
Back to top
View user's profile Send private message
onlinepancakes
Apprentice
Apprentice


Joined: 05 Sep 2006
Posts: 274
Location: Surprise - AZ

PostPosted: Sat Sep 16, 2006 2:53 am    Post subject: Reply with quote

I think it was good of them to mask 7.1 because of binary drivers. Would of blown hard if I went and did an emerge world and find out my damn drivers don't work and had to reinstall Xorg 7.0. Gentoo did their job perfectly by masking it for it wouldn't have fucked up on the 80% of us who's running Nvidia/ATI(AMD?) binary drivers.
_________________
Onlinepancakes --
Back to top
View user's profile Send private message
Slippery Jim
Apprentice
Apprentice


Joined: 08 Jan 2005
Posts: 264

PostPosted: Sat Sep 16, 2006 3:09 am    Post subject: Reply with quote

onlinepancakes wrote:
I think it was good of them to mask 7.1 because of binary drivers. Would of blown hard if I went and did an emerge world and find out my damn drivers don't work and had to reinstall Xorg 7.0. Gentoo did their job perfectly by masking it for it wouldn't have fucked up on the 80% of us who's running Nvidia/ATI(AMD?) binary drivers.


I agree, that would've sucked. (I use the ATI binaries).

It's better if portage detects when you have USE flags that bring in ati-drivers or nvidia-drivers, and does the right thing wrt which versions of xorg ebuilds to update.

So, for example, ati-drivers could require xorg <=7.0 as a dep, and you wouldn't get fragged. There's no reason we can't keep both versions of xorg in the stable branch, is there?
Back to top
View user's profile Send private message
boroshan
l33t
l33t


Joined: 16 Apr 2003
Posts: 730
Location: upside down

PostPosted: Sat Sep 16, 2006 8:45 am    Post subject: Reply with quote

Slippery Jim wrote:
NeddySeagoon wrote:
boroshan,

Rule 1. Do not feed the trolls
Rule 2. There is no rule 2 - see rule 1.


Ouch!

I wasn't trying to be a troll. Sorry if it sounded like I was.


It's okay, I think NeddySeagoon was trying to point out to me that I let myself get trolled by onanist wing of the FSF. I don't think he was getting at you.

Slippery Jim wrote:
I was trying to paraphrase the poll question, because it sounded to me like vipernicus read it differently than I did.


You're quite rght about the poll question too. I got merged in from another thread and keep forgetting that the title has changed on me.

NeddySeagoon wrote:
What are the lessons learned?

Personally? Next time, read the developer list archive first, so I know what's what from day one. Then, as Sam Vimes might have put it, if I still feel the need to go spare, I can do it with precision.

In the wider context, I'd like to think the lesson learned is if you have a package that is used near univerally, and that you know in advance is going to break the desktops of more than 50% of your userbase, make damn certain that the release mechanism works properly before you release it

Oh, and you're quite right about the benefit of hindsight; rule one compliance would get a lot simpler for a start.
_________________
Don't let THEM immanentize the Eschaton!
Back to top
View user's profile Send private message
Ibn al-Hazardous
Tux's lil' helper
Tux's lil' helper


Joined: 02 Sep 2004
Posts: 133
Location: Somewhere deep in the desert.

PostPosted: Sat Sep 16, 2006 10:13 am    Post subject: Reply with quote

Slippery Jim wrote:
onlinepancakes wrote:
I think it was good of them to mask 7.1 because of binary drivers. Would of blown hard if I went and did an emerge world and find out my damn drivers don't work and had to reinstall Xorg 7.0. Gentoo did their job perfectly by masking it for it wouldn't have fucked up on the 80% of us who's running Nvidia/ATI(AMD?) binary drivers.


I agree, that would've sucked. (I use the ATI binaries).

It's better if portage detects when you have USE flags that bring in ati-drivers or nvidia-drivers, and does the right thing wrt which versions of xorg ebuilds to update.

So, for example, ati-drivers could require xorg <=7.0 as a dep, and you wouldn't get fragged. There's no reason we can't keep both versions of xorg in the stable branch, is there?


This was what I meant with mutual blocks being good. You see, I was running xorg unstable to help finding packages that broke with modular when 7.1 entered unstable. And, to save the asses of us poor nvidia users they made 7.1 block nvidia drivers. All well and good?

Nope. You see nvidia drivers did not (originally) block xorg 7.1, so it happily installed itself on my 'puter. Before I found the cause, and masked and downgraded it I did a lot of hairpulling.

If blocks had been mutual, xorg 7.1 could have gone stable after a week for all that I care. And the poll misses out on an important alternative.
_________________
/Ibn
Back to top
View user's profile Send private message
slycordinator
Advocate
Advocate


Joined: 31 Jan 2004
Posts: 3065
Location: Korea

PostPosted: Sat Sep 16, 2006 11:12 am    Post subject: Reply with quote

I think he was saying that the decision to NOT stabilize was perfect.

Not that the method used was perfect.
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 54028
Location: 56N 3W

PostPosted: Sat Sep 16, 2006 5:40 pm    Post subject: Reply with quote

Slippery Jim,

My response on trolling was generic - when there is a troll around, ignore them, don't respond. It wasn't aimed at anyone in particular.

Whats not been mentioned in the thread so far, is that there were three issues, not only the two being discussed here.

There was
What to do about Xorg on x86 and amd64 in the face of the ABI change
What to do abount the binary drivers for Xorg.
What version of Xorg should be stable and used on the 2006.1 release media.

The decision to go stable with Xorg 7.0 and keep Xorg 7.1 in testing was made partly due to the (planned) timing of what was then the forthcoming 2006.1 release.

Xorg (well everything) on release media *must* be from the stable branch.
Xorg 7.1 would be difficult to make stable while there were no binary drivers. Stable users would need to mask it to get working binary drivers, which is something the stable users should never need to do.

With the timing and requirements of 2006.1 in the mix, on this occasion, keeping Xorg 7.0 as stable was probably the right decision. Just think of all the new users installing from 2006.1 being forced to learn about masking to use binary drivers. It would have put a '0' on the end of my post count.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3 ... 14, 15, 16
Page 16 of 16

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum