Canonical Releases "WLCS" Wayland Conformance Suite 1.0

[phoronix]

Phoronix: Canonical Releases "WLCS" Wayland Conformance Suite 1.0

While Ubuntu is not currently using Wayland by default with its GNOME Shell desktop and it doesn't look like they will try again until Ubuntu 20.10, the option is still available and they continue working in the direction of a Wayland Linux desktop future. One of their interesting "upstream" contributions in this area is with the Wayland Conformance Suite...

http://www.phoronix.com/scan.php?page=news_item&px=Wayland-Conformance-Suite-1.0

[ehansin]

Before anyone jumps in and says something along the lines of "Canonical should have supported Wayland from the start, Mir this, Mir that, blah, blah, blah..."

In my opinion, it was their money and resources, and they are entitled to "spend" them as they wish. And before you say it was all a waste of time, it is my opinion that sometimes you need to explore unknown territories before you can know if you are on the right path. And even if it turns out to be a dead end, often much is learned upon the journey.

I am not an Ubuntu fanboy or anything, I have played with many distros and love to explore and learn. Mark Shuttleworth, for as long as I remember, and to this day, has "here be dragons" on his personal blog. Maybe the guy likes to explore uncharted territory. And honestly, there were probably things learned from taking that divergent path. Maybe doesn't always seem helpful for the idea of a more cohesive Linux ecosystem, but things have seemed to come around, and things learned might just be of value as things move forward.

Again, just my opinion, and their resources, I have no right to tell them how to use them.

[r_a_trip]

In my opinion, it was their money and resources, and they are entitled to "spend" them as they wish. And before you say it was all a waste of time, it is my opinion that sometimes you need to explore unknown territories before you can know if you are on the right path. And even if it turns out to be a dead end, often much is learned upon the journey.

While this is true, it is also an oversimplification. The industry at the time had already put their blessing on Wayland as the way forward, including Canonical. Canonical did a 180 at the critical time where Wayland was about to anounce they had reached base protocol stability. At the time tiny Canonical thought they had clout enough to lead in the ecosystem and a lot of NIH projects came out of their stable, including Mir with a Canonical favouring CLA. Most of the Linux ecosystem received it with a "That's nice dear."

Years later, Mir and the vaunted converged Unity never made a dent. Unity is now shunted off to the community. Mir could luckily be salvaged as a Wayland display server. While predicting futures that never were are far from precise, I am willing to take a stab at it. Had Canonical put the same amount of effort in to making a Wayland compositor supporting Unity as a converged environment, we'd probably would have seen them lead with a seamless ecosystem between phone, tablet and desktop. Reinventing the wheel ultimately left Canonical a day late and a dollar short.

[renox]

Before anyone jumps in and says something along the lines of "Canonical should have supported Wayland from the start, Mir this, Mir that, blah, blah, blah..."
In my opinion, it was their money and resources, and they are entitled to "spend" them as they wish.

They are certainly free to spend their money as they wish, but the controversy was triggered by the blog which was "kind-of" criticising Wayland without saying really what was the issue.

And no, I'm not a fan of Wayland: if you want to have an efficient remote access IMHO text should be handled differently like X11 Render glyph cache..
Plus from a security point of view, SSD seems better than CSD (I know you can have both with Wayland but they used to push heavily CSD).

[r_a_trip]

I shouldn't, but I am going to.

*the recent clipboard feature for gnome3

So a clean up and reworking of an already present feature is reinventing the wheel?

*systemd

Bringing standardisation to the low level boot and service plumbing and bringing Linux more in line with the other Unices may be reinventing the wheel, but it is better to have round wheels than the triangular ones Linux had.

*pulseaudio

The low level sound architecture in Linux is an unholy mess. While PulseAudio had its teething problems, it certainly has made audio better manageable for a large number of people. It just works in most average cases. The same can't be said about pure ALSA.

*networkmanager

Same as with PulseAudio. While not 100% perfect, it has removed the need to fiddle with a lot of different tools for people who just want their network connection to work.

[ssokolow]

And no, I'm not a fan of Wayland: if you want to have an efficient remote access IMHO text should be handled differently like X11 Render glyph cache..
Plus from a security point of view, SSD seems better than CSD (I know you can have both with Wayland but they used to push heavily CSD).

I'll agree with you on SSD vs. CSD, but Wayland is just cutting the cruft out of what already developed when it comes to rendering. When modern versions of GTK+ and Qt use the MIT-SHM extension to talk to a compositing WM (the typical case these days), X11 has nothing to do with rendering except acting as "terrible, terrible IPC" because both the widget toolkits and compositors got fed up with X11's failings and they've already written in-house replacements for as much as they could under the X11 model.

I always point people to The Real Story Behind Wayland and X - Daniel Stone (linux.conf.au 2013) on this topic and it's where that "terrible, terrible IPC" phrase comes from. (It also goes into detail on why X11 has been "network capable, not network transparent" for ages.)

[cybertraveler]

Oooh, very nice. Things like this are very important for producing a stable platform for us all to develop on. Users will unknowingly benefit from efforts like this due to it creating greater compatibility between compositors.

[oiaohm]

IBM is reinventing the wheel, f.ex the recent clipboard feature for gnome3 (The computer clipboard was invented in seventies) . Systemd, pulseaudio and networkmanager are reinventing the wheel with poor design and quality too. This is the reason why Linux desktop is not more popular and IBM have a motive to do so. Everyone should use only IBM computers, f.ex an IBM server with thin clients. Gnome3 does a good job: resource hog, slow, buggy and keyboard focused.

Please be more careful you are committing legal deformation against IBM. Systemd, Pulseaudio and Networkmanager all started as work of Redhat.

Just because IBM acquired Redhat does not mean IBM is guilty of Redhat sins.

Pulseaudio was kind of reinventing the wheel but lets be realistic its a wheel rebuild that was required. Before Pulseaudio you have KDE with artsd , Gnome with esound this is only the start of path to 20+ different windows managers with own unique sound servers. Yes Pulseaudio was reinvention taking in all the features of all those different sound systems. Like pulseaudio network audio functionality does not come from esound or artsd it comes from "Network Audio System". Yes if you dig carefully though Pulseaudio you will find the markings of 20 different sound servers.

Yes Pulseaudio has come up short as it core design turned out not to be suitable to take in Jack-audio as well. Pulseaudio got us from 20+ sound servers down to 2 on Linux. Pulseaudio and jack-audio is now being lined up to be superseded by pipewire. Pipewire is almost the final step of get rid of sound servers. Keep on merging them until you have one left then you can work out kernel acceleration features.

Pulseaudio is better quality than the sound servers it replaced.

Networkmanager you could be right that it over stayed intel is working on connman as a competitor. This what was is networkmanger a reinvention of. Networkmanager started as the first graphical network manager on Linux. To be correct graphical is a stretch thinking it first form was text terminal but it was done with ncursors windows.. Yes nmtui is from 2004. Some of the problem is Networkmanager has basically the same interface layout it did when it first released in 2004 its been a functional design so it never evolved only been extended.

Give me a break on Clipboard was invented in the seventies. I wish it was just invented in the 70s and did not pick up the problems of the 70s. X11 and Textbased terminals each invented the clipboard solutions more than once. 1 on text based terminals is still alive today but in X11 there are 4 of them that you kind of have to be compatible with. So the clipboard was invented for Unix/Linux/BSD solutions that are still alive only 5 times with incompatibilities.

It would have been good if the Unix/Linux/BSD clipboard protocols was not out of the swinging 70s and being documented protocol mess like a 5 people on a drug fuelled bender trying to write a description of the round ball they are looking at and each one being used to make a ball. Of course the ball from this is going to come out all kind of malformed if you use those descriptions so is every single one of the Unix/Linux/BSD clipboard protocols so they all contain strange quirks. So expect gnome/kde to todo more clipboard solutions over time attempting to fix the god darn disaster.

Systemd hides a true horror story. Systemd starts merging in projects in the first place because as the lead developer of systemd is attempting to contact different projects to work with them find out hello this program no longer has a maintainer and in a few cases had not had a maintainer listening to bug reports for over 4 years. There has always been a question what happens if your lead developer gets hit by a bus and dies. Turns out for a lot of open source projects related to your system init process stuff like this had happened one of the projects merged into systemd the lead developer had been dead for 3 years before that and he was the only developer of that project. Turns out you get hit by a bus and die your project could and being called mature and everyone working around it quirks with distribution fixes instead of replacing the dead maintainer. Hate of systemd for merging things means people over look why it happened the story there is scary as hell.

Yes we had walking dead open source projects. Lucky where they were hosted was not depending on yearly or monthly payments or core projects to the init system of Linux could have completely disappeared. One way to reduce the odds of walking dead was to get projects with only 1 lead maintainer to merge under 1 big project so increasing number of maintainers with authorisation to repair and take over the projects as required.

This is why I really don't care that Systemd has upset stuff by merging stuff in or breaking people script based world. Long term Linux could be more durable with openrc, systemd, busybox and others doing their own implementations of different things. Even so the systemd merge has made the merge in projects more durable to maintainer loss.

[Buntolo]

Gnome3 does a good job: resource hog, slow, buggy and keyboard focused.

I agree on the first 3, but not on the 4th: I don't see how G3 is keyboard based. To me is horribly mice-based.

[nanonyme]

*systemd

Bringing standardisation to the low level boot and service plumbing and bringing Linux more in line with the other Unices may be reinventing the wheel, but it is better to have round wheels than the triangular ones Linux had.

I sometimes get the feeling the protest against systemd is mostly "at least these wheels are triangular and not square! we used to have a lot worse things before this".

When people think of reinventing wheels as a bad thing, they're going deep in the sunk-cost fallacy. Now as latest evolutionary step of sound servers Pipewire might merge PulseAudio and Jack userbases as well.