Ubuntu 20.10 Released With GNOME 3.38, Active Directory Installer Integration

[JackLilhammers]

I have a RX 560 on amdgpu.
Distros are Ubuntu and Manjaro (and please don't even say these are crap).
Situations:
4k30 X: OK
4K30 Wayland: OK
4K60 X: OK
4K60 Wayland: NOK (playing at twice - 2x 30? - the speed and distorted high-pitched voice)

This was just a blocking example. There are many other issues. Not just for me but everywhere you read about wayland, you see people explaining stuff that doesn't work for them. I don't want to discuss it further either beyond that example, I just mentioned to illustrate the limited perspective of some cultists.

I think that most wayland cultists base their opinion on their limited experience and their limited use case.
Which would be fine to show that there is a bug, but is not enough to say that the whole platform works.
From the user's perspective the transition from xorg to wayland should be seamless and transparent. One should not even notice.
Anything less would be unacceptable on any other major platform.

PS: As a developer I can understand the process, the pita that is working with Nvidia, the fact that many use cases work, etc.
But as a simple user I couldn't care less about anything that has to be done to have Nvidia graphic cards up and running on par (ish) with Windows.
Until then, Wayland is broken.

[mystiquewolf]

You can think of it like Windows Apps, or like Play Store. Play Store is not bad IMO. The difference is in the number of available apps.

[dragon321]

Not just for me but everywhere you read about wayland, you see people explaining stuff that doesn't work for them. I don't want to discuss it further either beyond that example, I just mentioned to illustrate the limited perspective of some cultists.

You can also find complains about bad experiences with HiDPI support on Xorg (especially with mixed DPI configurations) and there are still voices that Wayland is broken, unneeded and useless because "we have Xorg". What advice you gonna give to these peoples? Maybe "Use Windows or macOS"? Well, things like that works fine on these operating systems. Would you like to call Xorg "broken"?

I believe it's the same kind of crap harmful as a whole to the Linux ecosystem.

So you think Linux ecosystem should stay fragmented and with broken display? Should be worse desktop platform than Windows or macOS that nobody cares about? This is what you want for desktop Linux?

[Mez']

You can also find complains about bad experiences with HiDPI support on Xorg (especially with mixed DPI configurations) and there are still voices that Wayland is broken, unneeded and useless because "we have Xorg". What advice you gonna give to these peoples? Maybe "Use Windows or macOS"? Well, things like that works fine on these operating systems. Would you like to call Xorg "broken"?

Xorg works better in the multitude of use cases out there. Doesn't mean it's perfect or it shouldn't be replaced.
Because, yes, wayland will replace it someday. I believe though that people and distros should wait a couple more years (my advice), for early adopters (from Fedora for example) to report bugs and wayland devs to iron them out. If it's as ready as they claim (for their few use cases), let them get the crippled experience first.
And let the others opt for Wayland by default when it has matured enough.

It might be a pain in some scenarios though. For example, Gnome is acting up or eating up unreasonable memory very often due to its flawed design. And not being able to alt + f2 for a quick refresh (but log out and back in instead losing all your current open stuff) is going to be a huge downgrade.

So you think Linux ecosystem should stay fragmented and with broken display? Should be worse desktop platform than Windows or macOS that nobody cares about? This is what you want for desktop Linux?

I was talking about snap/flatpak here, not about display.
And yes, I'm fine with how packages are handled now. It's just amazing. I'd be alright with a common package management with shared libraries (deb/rpm style), never for redundancy of bundled libraries, suboptimal mutualization of resources, inconsistent theming and other borked design principles of flatsnapcrapimage. Clearly.

[CochainComplex]

I was addressing Volta.
I'm just tired of the Ubuntu bashing from certain Fedora users blatantly disregarding the many wayland bugs and the diversity of workflows that vanilla Gnome doesn't see to, or promoting flatpaks while I believe it's the same kind of crap harmful as a whole to the Linux ecosystem.
Blah blah everything Fedora does is better and the single version of the truth, blah blah Ubuntu is wrong. They sound like SJWs/liberals/regressists in their one-track thinking. Instead of tolerating others' ways.

...I see sorry my fault. Just take out the personal arguments

[CochainComplex]

I have a RX 560 on amdgpu.
Distros are Ubuntu and Manjaro (and please don't even say these are crap).
Situations:
4k30 X: OK
4K30 Wayland: OK
4K60 X: OK
4K60 Wayland: NOK (playing at twice - 2x 30? - the speed and distorted high-pitched voice)

This was just a blocking example. There are many other issues. Not just for me but everywhere you read about wayland, you see people explaining stuff that doesn't work for them. I don't want to discuss it further either beyond that example, I just mentioned to illustrate the limited perspective of some cultists.

Ok valid argument. I still think the major issue is that even if Wayland is dedicated to be the successor of X the majority of devs are treating first X then Wayland. As said before there are many argument of pro and cons and both sides have good arguments but we have this orobus situation.

If Wayland gets not defaulted it will not have a splendid userexperience but this very fact is blocking the default adaptation. Pushing it now to a widely adopted Distribution like Ubuntu will really push devs to iron out the issues.

e.g. Prime iGPU GPU handling is (was) a real pain.. if you have the classic distros but it is super easy and comfortable with PopOS ...devteam is not huge. resources are far behind canonical or red hat..but still they seem to get it right. why? Because they have decided it is one of their priority to make this as easy to use as possible.

For X vs Wayland ...Most distros: Well here is gnome, kde on x we have xy Bugs with ..we need to iron it out for next release. and then on top we have xy wayland bugs. Fix X first then the wayland things... Releasedate: we have fixed 230 of 300 X bugs but 50 of400 Wayland bugs ...and this is the case from release to release to release.

[dragon321]

Xorg works better in the multitude of use cases out there. Doesn't mean it's perfect or it shouldn't be replaced.
Because, yes, wayland will replace it someday. I believe though that people and distros should wait a couple more years (my advice), for early adopters (from Fedora for example) to report bugs and wayland devs to iron them out. If it's as ready as they claim (for their few use cases), let them get the crippled experience first.
And let the others opt for Wayland by default when it has matured enough.

It might be a pain in some scenarios though. For example, Gnome is acting up or eating up unreasonable memory very often due to its flawed design. And not being able to alt + f2 for a quick refresh (but log out and back in instead losing all your current open stuff) is going to be a huge downgrade.

Didn't we wait enough time? Wayland protocol is stable since version 1.2 released in 2013. It's seven years. Majority of Linux users should already use Wayland but they are not. It's about time to promote Wayland as X11 replacement and start using it. Setting is as default choice with Xorg as fallback option is one of the way to promote it. If Xorg will stay default everywhere then why developers (of desktop environments, applications etc.) would even care about Wayland at all? Simply telling them "Care about it because it's future" won't work. Actual users and their requirements will.

I was talking about snap/flatpak here, not about display.
And yes, I'm fine with how packages are handled now. It's just amazing. I'd be alright with a common package management with shared libraries (deb/rpm style), never for redundancy of bundled libraries, suboptimal mutualization of resources, inconsistent theming and other borked design principles of flatsnapcrapimage. Clearly.

Just like I did. Whether you like it or not Linux desktop is fragmented and making software release is way more annoying than it is for Windows or macOS. Not only we have many package formats but also package made for distribution XYZ 1.0 can break in XYZ 2.0 because some dependencies have changed. You solved this? Ok, then do it again with distribution ZYX which uses different package format. Practically the best way to make sure your package won't break between releases is to bundle it with needed libraries like Steam or games doing. Should distributions agree on common standard? Well, we had LSB which failed. Even relying on distributions repositories maintainers is not solving everything. First - it causes delay and some packages in repositories can stay outdated even for several months. Second - they won't cover any application, especially closed source ones. Then you look at Windows or macOS and these problems simply don't exists here.

Containers provides a way to make one package which should work on most distributions. Yeah, there are some problems with integration but it's a small price for reducing fragmentation. Also themes works with containers, that problem was solved years ago. Resources? More disk usage is another small price for making developers lives easier. Using Flatpak doesn't mean you need to have many duplicated libraries. Of course it's not perfect but it's needed. It wouldn't be if fragmentation wouldn't exist. But it exists.

[Mez']

Didn't we wait enough time? Wayland protocol is stable since version 1.2 released in 2013. It's seven years. Majority of Linux users should already use Wayland but they are not. It's about time to promote Wayland as X11 replacement and start using it. Setting is as default choice with Xorg as fallback option is one of the way to promote it. If Xorg will stay default everywhere then why developers (of desktop environments, applications etc.) would even care about Wayland at all? Simply telling them "Care about it because it's future" won't work. Actual users and their requirements will.

Because wayland is coming. It's a fact. Even though it will make some lives harder (the refresh I mentioned). But also some other lives easier. Developers are on it. There's a whole article about Xorg being abandonware on this very site.
Doesn't mean we have to make wayland default when it doesn't yet work properly. I can't, I have at least 3 blocking bugs, and 5 light annoyances due to it. When you read about wayland, it's always "Everyone should go on it already, it works. I only have that one issue/bug/annoyance..." (adding them up makes for the yellow pages). Especially on Ubuntu where there are far more use cases and workflows (and thus extensions to address them) used than in Fedora.
People won't switch just for the sake of switching and enduring bugs and issues. They will switch when using doesn't bother them in any single way.

Just like I did. Whether you like it or not Linux desktop is fragmented and making software release is way more annoying than it is for Windows or macOS. Not only we have many package formats but also package made for distribution XYZ 1.0 can break in XYZ 2.0 because some dependencies have changed. You solved this? Ok, then do it again with distribution ZYX which uses different package format. Practically the best way to make sure your package won't break between releases is to bundle it with needed libraries like Steam or games doing. Should distributions agree on common standard? Well, we had LSB which failed. Even relying on distributions repositories maintainers is not solving everything. First - it causes delay and some packages in repositories can stay outdated even for several months. Second - they won't cover any application, especially closed source ones. Then you look at Windows or macOS and these problems simply don't exists here.

Containers provides a way to make one package which should work on most distributions. Yeah, there are some problems with integration but it's a small price for reducing fragmentation. Also themes works with containers, that problem was solved years ago. Resources? More disk usage is another small price for making developers lives easier. Using Flatpak doesn't mean you need to have many duplicated libraries. Of course it's not perfect but it's needed. It wouldn't be if fragmentation wouldn't exist. But it exists.

I still disagree, sorry. Yes, it requires more effort. But that effort is completely worth it and makes for the greatness of Linux in my opinion. Downgrading to bundled packages is doing what Windows does worse. It's a big loss and one less argument for trying Linux over Windows. You get the same wrong use of ressources.
I'm open to snapflatcrap for entry-level proprietary stuff (closed source packages). For easily porting apps. But just as a transitory period, and move on to a more solid and thought out design when they know the ecosystem.
Also theming has not been solved. There are still a lot of inconsistencies depending on the app and the availability or customization of themes. Oh, right... In Fedora, theming is discouraged and almost forbidden as it requires an extension (you know... vanilla Gnome and all). Probably why they don't see it.
Finally, yes it means duplicated libraries. They specifically said that flatpak (but same for snaps) will only reuse the same library when an app uses a version of the library already bundled. As inevitably devs or editors will quickly all do their own thing (devs won't even care if they are allowed to do what they want), and apps will include a lot of different libraries version depending on the moment of release, the frequency of the release (older stuff for example won't just change packages to refresh libraries), etc... It means very little reuse over time (I'm talking over 2-3 yaers) and a lot of fragmentation and resources wasted.

[dragon321]

Because wayland is coming. It's a fact. Even though it will make some lives harder (the refresh I mentioned). But also some other lives easier. Developers are on it. There's a whole article about Xorg being abandonware on this very site.
Doesn't mean we have to make wayland default when it doesn't yet work properly. I can't, I have at least 3 blocking bugs, and 5 light annoyances due to it. When you read about wayland, it's always "Everyone should go on it already, it works. I only have that one issue/bug/annoyance..." (adding them up makes for the yellow pages). Especially on Ubuntu where there are far more use cases and workflows (and thus extensions to address them) used than in Fedora.
People won't switch just for the sake of switching and enduring bugs and issues. They will switch when using doesn't bother them in any single way.

It's coming something like 10 years (or even more) and it's still not default choice on most distributions. As I said it's about time to make it default. Your blockers won't affect everybody. Some users will be fine with Wayland and rising popularity will make developers care about it.

I still disagree, sorry. Yes, it requires more effort. But that effort is completely worth it and makes for the greatness of Linux in my opinion. Downgrading to bundled packages is doing what Windows does worse. It's a big loss and one less argument for trying Linux over Windows. You get the same wrong use of ressources.
I'm open to snapflatcrap for entry-level proprietary stuff (closed source packages). For easily porting apps. But just as a transitory period, and move on to a more solid and thought out design when they know the ecosystem.
Also theming has not been solved. There are still a lot of inconsistencies depending on the app and the availability or customization of themes. Oh, right... In Fedora, theming is discouraged and almost forbidden as it requires an extension (you know... vanilla Gnome and all). Probably why they don't see it.
Finally, yes it means duplicated libraries. They specifically said that flatpak (but same for snaps) will only reuse the same library when an app uses a version of the library already bundled. As inevitably devs or editors will quickly all do their own thing (devs won't even care if they are allowed to do what they want), and apps will include a lot of different libraries version depending on the moment of release, the frequency of the release (older stuff for example won't just change packages to refresh libraries), etc... It means very little reuse over time (I'm talking over 2-3 yaers) and a lot of fragmentation and resources wasted.

Maybe worth for you but it doesn't look like it's worth for many developers. They don't want to spend time making packages just because its "better". Yeah, packages would be ideal solution without fragmentation but as I said fragmentation exists and packages pros aren't that great when you have to deal with cons. As long making package for Linux won't be easy as for macOS and Windows "snapflatcrap" will be needed unless you don't want to popularize Linux and keep it in ~3% forever. Also Linux containers are not identical to Windows. There are some differences and they are better solution. For example permission control is implemented better on Flatpak.

Theming was solved, you can easily install theme in Flatpak and make applications use it. It's working so how it's not solved? Nobody also said it's forbidden and you shouldn't use it. Neither Flatpak or Fedora forbids theming. Yeah, staying with default theme is easiest way but nobody forces you to do it. And for peoples like you Flatpak got theming support.

Flatpaks have runtimes just for that. Runtimes contain basic libraries used by every app that is build for specified firmware. Application only contain some less common libraries which usually don't take up much space. Updating runtime updates every application just like with package managers. Even if there are duplicated libraries you can use deduplication if you really care about more megabytes disk space in 2020. Yes, they will take more space but it's worth price. Especially with things like deduplication or today's disk sizes.

[Mez']

It's coming something like 10 years (or even more) and it's still not default choice on most distributions. As I said it's about time to make it default. Your blockers won't affect everybody. Some users will be fine with Wayland and rising popularity will make developers care about it.

I'm not talking about just the bugs I encounter. Whenever you read a forum on the topic, everyone goes with his own issue. When piling up, there are dozens of them.

Maybe worth for you but it doesn't look like it's worth for many developers. They don't want to spend time making packages just because its "better". Yeah, packages would be ideal solution without fragmentation but as I said fragmentation exists and packages pros aren't that great when you have to deal with cons. As long making package for Linux won't be easy as for macOS and Windows "snapflatcrap" will be needed unless you don't want to popularize Linux and keep it in ~3% forever. Also Linux containers are not identical to Windows. There are some differences and they are better solution. For example permission control is implemented better on Flatpak.

I care about developers to a limited extent. I've always been on the user side, due to my profession. The problem of Gnome, for example, is that developers have thought the design for themselves and then they pushed it to users. Here, take this blob stripped down to the minimum and we don't care if it doesn't fit your workflows. Just change your workflow and shut up. This is the wrong approach and it has shown given how Gnome is criticized and loved-hated, even by its own users.
Systems need to be thought for the end users, to adapt to its various workflows and use cases. That's why Unity will always be more popular with Ubuntu users, because it had them in mind and could be tailored to anything you needed. And the user wants that.
Here, again, flatpaks (or snaps) have just been created for the devs conveniency. But for the user, even if it's becoming a bit more transparent, it's still an annoyance and requires more effort for download and maintenance.

Theming was solved, you can easily install theme in Flatpak and make applications use it. It's working so how it's not solved? Nobody also said it's forbidden and you shouldn't use it. Neither Flatpak or Fedora forbids theming. Yeah, staying with default theme is easiest way but nobody forces you to do it. And for peoples like you Flatpak got theming support.

The typical horse blinkers due to not taking various workflows and preferences into account.
Theming it not solved. If you want a consistent theme, you need to stick to themes available in flatpaks or snaps, you can't just take any theme, you are limited to what is included. Also, I customize my themes. Even if stock theme was available, it would still be inconsistent. So, I repeat, theming is absolutely not solved except for the people with very limited use cases.
Regarding the default theme being discouraged or forbidden, it was a tease. Fedora cultists often insist vanilla Gnome is the single version of the truth, and everyone should use it exactly as designed regardless of personal conveniences (since they don't care about preferred personal workflows). Hence, vanilla Gnome = no user themes extension = Adwaita.
But I'm pretty sure in the end, since Fedora users have restrained workflows, they always think something works. Reality is, the moment it gets out to other distros with much wider and general use cases, it breaks from everywhere and exposes tons of bufs. That's what happens with wayland to come back to why it can't be default yet.

Flatpaks have runtimes just for that. Runtimes contain basic libraries used by every app that is build for specified firmware. Application only contain some less common libraries which usually don't take up much space. Updating runtime updates every application just like with package managers. Even if there are duplicated libraries you can use deduplication if you really care about more megabytes disk space in 2020. Yes, they will take more space but it's worth price. Especially with things like deduplication or today's disk sizes.

It's never been a question of megabytes disk space (although I have some limited partitions when trying out new distros), it's a question of mutualization, redundancy, optimization, streamlining, which helps using less ressources and get faster systems. You don't have to design badly and heavy because you have the space, size, money, time, whatever, for it. There's always a price to pay for that. And I don't want to pay it.
Plus I just really don't like how flatpaks and snaps are managed, it's like going back to the middle age. It reminds me of the terrible days of Windows 95 or 98.