X.Org Server 1.18 Officially Released

[omer666]

Right, except that invert the reality is always an hard activity, isn't it?
Nobody can target Fedora 23 at launch, Nvidia included, because they have shipped a development version of x.org server.
x.org is not closed source software, so they were well informed that the 1.18 would not be released in time for F23. So they have deliberately included a development version of one of the most important package, then your "invert the reality" activity should just stop here.

It is so hard an activity that I don't even consider trying to do it. In fact I wouldn't even know what an "inverted reality" could possibly look like.
Fedora 23 has been out for one week before the 1.18 final release, so considering Fedora's release cycle, I am not really convinced it was such a mistake.

Also as you point out yourself, Xorg is open source so nVidia could just get their hands on the new API and update their drivers, or at least try to do so for the stable Xorg release.

In an open source ecosystem, closed source software should just adapt to the system. As I said, they just chose not to do so and I am fine with it. Just don't try to accuse Fedora of messing up because of some (temporarily) incompatible proprietary software. They don't care and they don't have to.

I don't know what you mean by "inverting reality", but I sure know what denial is...

[bobwya]

Nvidia will not be in good shape... Still waiting for them to pull their finger out of their ass and fix the rather majorly annoying DisplayPort issues with the newer Nvidia driver versions (>352.30)...

When your laptop display goes into DPMS sleep and doesn't wake again... Yeah that's not very useful...

[valeriodean]

It is so hard an activity that I don't even consider trying to do it. In fact I wouldn't even know what an "inverted reality" could possibly look like.
Fedora 23 has been out for one week before the 1.18 final release, so considering Fedora's release cycle, I am not really convinced it was such a mistake.

Also as you point out yourself, Xorg is open source so nVidia could just get their hands on the new API and update their drivers, or at least try to do so for the stable Xorg release.

In an open source ecosystem, closed source software should just adapt to the system. As I said, they just chose not to do so and I am fine with it. Just don't try to accuse Fedora of messing up because of some (temporarily) incompatible proprietary software. They don't care and they don't have to.

I don't know what you mean by "inverting reality", but I sure know what denial is...

Thank you, but I don't need another reality inverted example. :-)

1) release a stable version of the distro with the x.org server in dev version just to save a week sounds really stupid to me. Even more stupid because waiting for a week would be saved them from this ugly showing, for free.
2)

nVidia could just get their hands on the new API and update their drivers, or at least try to do so for the stable Xorg release

That happens after the F23 release. :-)
So even in case of zero delay by nvidia, still you had no driver available in time for the release day. And the day after. And the day after...
3)

They don't care and they don't have to.

Of course, because they made fedora just to show to themselves that they are capable to compile some source code and then assembly a linux distro.
The users and their needs are just a strange side effect that they tolerate with patience.

Ah, just to inform you: the nouveau driver are not stable as you like to think. Not a single day goes without graphical corruption and, after some seconds, the desktop freeze to the point that I must do an hard reset of the pc, without a proper shutdown as a consequence.
This behaviour started after the switch to nouveau and repeat it every day.
I have an idea for the next codename: F24 - for_intel_user_only

There is something that they can do for F24: remove the video driver blob before the upgrade, because makes absolutely no sense leave the blob installed by the user to the cost of a failed boot.
In this way the user will have a bootable system (in any cases, not matter how bleeding edge is the new kernel, the new mesa, the new x.org or whatever) and the fedora devs will not be responsible for an upgrade that resulting in a black screen as a not-so-happy end.

[airlied]

Thank you, but I don't need another reality inverted example. :-)

1) release a stable version of the distro with the x.org server in dev version just to save a week sounds really stupid to me. Even more stupid because waiting for a week would be saved them from this ugly showing, for free.
2)
That happens after the F23 release. :-)
So even in case of zero delay by nvidia, still you had no driver available in time for the release day. And the day after. And the day after...
3)
Of course, because they made fedora just to show to themselves that they are capable to compile some source code and then assembly a linux distro.
The users and their needs are just a strange side effect that they tolerate with patience.

Ah, just to inform you: the nouveau driver are not stable as you like to think. Not a single day goes without graphical corruption and, after some seconds, the desktop freeze to the point that I must do an hard reset of the pc, without a proper shutdown as a consequence.
This behaviour started after the switch to nouveau and repeat it every day.
I have an idea for the next codename: F24 - for_intel_user_only

There is something that they can do for F24: remove the video driver blob before the upgrade, because makes absolutely no sense leave the blob installed by the user to the cost of a failed boot.
In this way the user will have a bootable system (in any cases, not matter how bleeding edge is the new kernel, the new mesa, the new x.org or whatever) and the fedora devs will not be responsible for an upgrade that resulting in a black screen as a not-so-happy end.

Stockholm Syndrome must be hard. Aaron from nvidia is also doing some truth avoidance techniques that obviously work on you.

the ABI for 1.18 froze at RC1 which was released on the 1st September. Aaron is the one who generally makes sure we stick to upstream ABI freezes, so I'm surprised when he tries to use the fact the final release hadn't happened as an excuse. Maybe we shouldn't bother with upstream ABI freezes anymore.

Also it was one of the Fedora X maintainers that actually did the rc2 and final X server releases upstream, he'd have done them earlier but schedules were all over the place and the 1.18 X schedule was a disaster.

Fedora 23 had X 1.18 in it because X 1.18 has features that the open source users want, and closed source users don't have to upgrade to F23 on release day they can wait a month or two for their selected driver maintainers to catch up.

But there is no way we are holding upstream X or Fedora hostage to nvidia/fglrx.

Dave.

[darkbasic]

Stockholm Syndrome must be hard. Aaron from nvidia is also doing some truth avoidance techniques that obviously work on you.

the ABI for 1.18 froze at RC1 which was released on the 1st September. Aaron is the one who generally makes sure we stick to upstream ABI freezes, so I'm surprised when he tries to use the fact the final release hadn't happened as an excuse. Maybe we shouldn't bother with upstream ABI freezes anymore.

Also it was one of the Fedora X maintainers that actually did the rc2 and final X server releases upstream, he'd have done them earlier but schedules were all over the place and the 1.18 X schedule was a disaster.

Fedora 23 had X 1.18 in it because X 1.18 has features that the open source users want, and closed source users don't have to upgrade to F23 on release day they can wait a month or two for their selected driver maintainers to catch up.

But there is no way we are holding upstream X or Fedora hostage to nvidia/fglrx.

Dave.

Liking the post is definitely not enough, I completely agree.

[pal666]

well, the Nvidia blob is certainly not ready yet for 1.18. Fedora 23 shipped with a pre-release of xorg-server 1.18 and the drivers are certainly not compatible yet.

it is not for the first time. but some people still don't understand that nvidia has no linux driver

[pal666]

Nobody can target Fedora 23 at launch, Nvidia included, because they have shipped a development version of x.org server.

utter bullshit
everyone can target fedora 23, except proprietary shit-eaters, but that is their own problem. mesa drivers work wonderfully in f23

[kmare]

it is not for the first time. but some people still don't understand that nvidia has no linux driver

yeah, to be honest I upgrade to f23 without knowing there was no nvidia driver (well, actually I thought that xorg-server was still at 1.17) so now I have a pc without being able to use the blob. As I mainly do android and web development, right now it's not a problem. But from time to time I do some 3D development and at the moment I can't really use nouveau for that (although it works wonderfully for everything else - don't really do gaming). I really hope nvidia management will wake up at some point and start investing in the open source drivers like the rest (or most of them).

[pal666]

Just read phoronix tests, proprietary drivers have better and more solid fps.

just read this thread, proprietary windows shit does not work under linux in general

Gamers do not touch to oss display drivers if possible. I am sure that nvidia 358 driver from nvidia site works with latest fedora, because it works fine with debian testing and kernel 4.3.0. In debian the packaged drivers not not work or are old. It is like in windows, get your display driver from manufacturer site.

linux gamers do not use nvidia because nvidia has no linux driver(windows shit in linux wrapper, which does not really work, does not count). there are enough games which work nice with radeonsi

[valeriodean]

Stockholm Syndrome must be hard. Aaron from nvidia is also doing some truth avoidance techniques that obviously work on you.

the ABI for 1.18 froze at RC1 which was released on the 1st September. Aaron is the one who generally makes sure we stick to upstream ABI freezes, so I'm surprised when he tries to use the fact the final release hadn't happened as an excuse. Maybe we shouldn't bother with upstream ABI freezes anymore.

Also it was one of the Fedora X maintainers that actually did the rc2 and final X server releases upstream, he'd have done them earlier but schedules were all over the place and the 1.18 X schedule was a disaster.

Fedora 23 had X 1.18 in it because X 1.18 has features that the open source users want, and closed source users don't have to upgrade to F23 on release day they can wait a month or two for their selected driver maintainers to catch up.

But there is no way we are holding upstream X or Fedora hostage to nvidia/fglrx.

Dave.

I don't know who is Aaron and I don't care, but for sure you can leave the Syndrome of Bla Bla Bla to yourself. What I described is what happened to my pc, and that cannot be influenced by the Aaron's opinion neither in a way nor in another.
So the fedora x maintainers taken care of the x 1.18 releas? Good to know, but then: why you failed so hard to manage the release of both?
Let me summarise:
1) fedora devs managed the F23 release
2) fedora x mainteiners managed the x.org 1.18 release
3) F23 was shipped one week before the x release
Do you recognise nothing a bit strange in the bullet points above? Nothing?
For the next time, my lord, speak less about the Stockholm Syndrome and work better no the release schedule.
I saw x.org version 1.17.99 (or 1.17.90, don't remeber) after F23 installation, and this means a development version of x.org, and this means you shipped a (6 month) distro release with a development version of the server X.
Now try to apply some of your technique to avoid this truth.

Oh, yeah, speaking about "features that the open source users want", now that I'm a (forced) open source user , add the following point to the list: a damn system that doesn't freeze two times in a day, thanks.

That being sayd, thank you devs for your work. It is just beyond me why you love so much get lost in a glass of water.

Ah, just the last very important thing: why do you leave the closed driver installed during the distro upgrade? It is a 100% fail scenario, so why?