Announcement

Collapse
No announcement yet.

KDE Saw More Wayland Fixes This Week, Other Changes

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Originally posted by oiaohm View Post

    You did not get what I meant by being careful what you ask for.
    1) People ask for more control over a device this results in coding the GUI to give that more control is now more complex to make because you have more controls.
    2) Working out where you should store all those extra options is another level of complexity this has changed 3 times in libinput history.
    3) libinput is under active development so new features are being added.

    Humble users like you need to be aware that synaptics
    1) Has decent GUIs now because its development is dead and is really feature poor but when synaptics was under active development it was quite common to have no GUI like libinput now.
    2) That at some point synaptics will no longer be usable. That will either because we move to wayland or the Linux/BSD kernel kill synaptics interface. So synaptics on borrowed time.

    Maybe you want to be involved in the libinput process talking about what control features you want exposed and where. Maybe you want to do a quirk file for you device and submit it so out the box works decently with libinput. Remember the developer of libinput cannot own every version of a touch pad how can the libinput developer know that your model touchpad is problem if you don't report it.

    Like it or not this libinput problem is not a wayland only problem. libinput problems is down road for the X11 server on bare metal for all those still using synaptics as well if X11 server on bare metal remains in use long enough. The work around of synaptics driver with x.org on bare metal is not a forever solution fixing libinput so your touchpad works right is the forever solution.
    I would say that this fully encompasses my thoughts.

    Comment


    • #32
      Originally posted by Charlie68 View Post
      I would say that this fully encompasses my thoughts.
      https://www.askwoody.com/forums/topi...input-drivers/
      There is a big mistake there that really important. "synaptics" driver is not only for synaptics branded touchpads is true but its also synaptics standard touchpads. synaptics use to hold a patent that is now expired any party who bought that patent agreed to make their touchpads to a particular quality and standard. Currently have the attack of the badly made clones. So there are touchpads that you can attempt to use the synaptic X11 driver with and end up in kernel panic hell.

      Lot of people have not noticed how many windows laptops have custom touchpad drivers what are fairly much just the same driver as all their other models with quirks/profile data for that model.

      The general setting you get with windows are already behind quirk/profile data. Something to wake up here that is big problem. People are buying lots laptops for Linux lots vendors don't submit anything to libinput. Remember in the time synaptics driver was in active development those companies who bought patents from synaptics was submitting profile data on the devices they were making back to synaptic this information did go to the developer making the driver for X11.

      Yes I know its horrible that end user has to install diagnostic tool and collect profile data. Let be realistic here the libinput developers don't have patents to demand the information for the device makers. So if they don't have end user run diagnostic tools how else are they going to get it as they cannot afford to go out and buy every one.

      This is some of the problem as well not understanding that there is a lot of problems here.

      I do think the diagnostic tool could be made more GUI. But setting up a touchpad that does not have a quirks/profile file is going to remain painful.

      Comment


      • #33
        On my KDE Debian, lininput is an installed dependency and synaptics is not installed. That's with default settings out of the box.
        So from my point of view synaptics is already off the screen if not installed manually.
        It also gave me problems inside virtual machines so (as an end-user) I'm not too sad about that direction, for me it just has to work.

        Comment


        • #34
          Originally posted by Charlie68 View Post

          I would say that this fully encompasses my thoughts.
          https://www.askwoody.com/forums/topi...input-drivers/
          Please do note there is only 1 core developer


          I am serous all the work of Libinput is 90%+ 1 developer that we can thank out lucky stars is funded by Redhat on one hand. On another hand enterprise deployments unique mouse/trackpad configuration per user is not ideal either.

          For such a important thing synaptics has no developer. The backend library that evdev depends on in X11 x.org bare metal is also 90% the work of the same developer Peter Hutterer.

          Can you now see why Libinput does not have a decent GUI. Simple reality with the number of devices libinput supports for single developer there is no enough hours in the day for that developer to go make a proper GUI for libinput. There is not even enough hours in the day todo all the device support work need.

          Please note Peter Hutterer is not qualified in UX/GUI design either. So libinput getting a decent UI/GUI really requires a new developer.

          If this was like the Linux kernel you would be expecting you look at the contributors and you would be seeing dell/hp.... all the different laptop vendors with touch pads there with a developer providing assistance so Linux worked well on their hardware. Guess what they are not there.

          We should be thankful that libinput is as good as what it is with how under resourced the project is. Yes in the past you got spoiled rotten by a well funded project that was the synaptics driver.

          Yes the lead developer saying he has to restrict features because he does not have the means to test it this was a smoke signal about being under resourced.

          Please note I am not saying libinput cannot be better. But we have to be real here libinput project is serous-ally lacking resources. And the synaptics driver is critically lacking resources to the point its going to come non functional.

          Comment


          • #35
            hmm, interesting, I only use libinput on a few machines > Kbuntu 20.04 + our optimizations and it works very well - an example from Lenovo 720s-IKB [14" + MX150]:

            https://www.dropbox.com/s/i35mxahc47...90643.png?dl=0
            Last edited by ext73; 21 March 2021, 04:11 AM.

            Comment


            • #36
              Ok, but we do not compare to Windows, where thanks to a gui you can configure it as you want.
              I'm not saying libinput has bad configuration on all devices, but it's just this stupid idea that needs to work with default settings that doesn't make sense.
              Because each of us has their own preferences, especially on a toucphad where the sensitivity changes from person to person. To think that there is a method to standardize everything is just stupid.
              It is not a question of how many people are working on libinput, but it is the very basic idea that is wrong.
              My God ! We are in 2021, we are full of useless DE with special effects and still asking people to debug and diagnose to optimize a touchpad?
              The problem is that they don't want to do it and not because there is no one who can do it, but because they say, the less complicated it is the less bugs there will be and the less work to maintain it.
              This is just unacceptable and reminds me a lot of the logic behind a particular DE.

              Comment


              • #37
                Originally posted by Charlie68 View Post
                I'm not saying libinput has bad configuration on all devices, but it's just this stupid idea that needs to work with default settings that doesn't make sense..
                This is kind of a mistake. If you survey MS windows and Mac OS users 90%+ of them have never touched mouse/touchpad settings and don't even know where it is.

                Originally posted by Charlie68 View Post
                Because each of us has their own preferences, especially on a toucphad where the sensitivity changes from person to person. To think that there is a method to standardize everything is just stupid.
                This is not true. The reality is majority of users don't care as long as its in the ball park its good enough they don't change the sensitivity to suite them.

                Out the box experience is important for 90%+ of users. Yes you are in the less than 10% of users who care and want to adjust mouse and pointer setting at all.

                Originally posted by Charlie68 View Post
                It is not a question of how many people are working on libinput, but it is the very basic idea that is wrong.
                How under resourced is a fact. This does effect how much the developer has to focus on market share and what is possible.

                Originally posted by Charlie68 View Post
                My God ! We are in 2021, we are full of useless DE with special effects and still asking people to debug and diagnose to optimize a touchpad?
                Yes for touchpads that libinput does not in fact have profile for this is required. You will notice quite a few users with particular brands say libinput is decent out the box because the vendor did the profiling for them but this is only 3 brands of laptops.

                Originally posted by Charlie68 View Post
                The problem is that they don't want to do it and not because there is no one who can do it, but because they say, the less complicated it is the less bugs there will be and the less work to maintain it.
                This is you miss reading the statement. There is not enough resources so now the lead maintainer has to optimise. Less complicated, Less bugs and suites 90% of users to be that way and fits with the amount of resources libinput has.

                The reality you want more to handle more complicated with more work to maintain it you need more developers.

                The reality is they don't want to do it arguement has you failing to step back look at the big picture see that you are in the minority and the developer is under resourced so has to make hard choices. Just like libinput does not come with a decent GUI program as part of libinput this is resources. Think about it what is the point about giving users mouse curve control if they don't have a good GUI to set it valves and diagnose if it right. Lack of GUI developer has knock on effects.

                The reality here you are trying to get 1L of water out of a cup of water. If the resources as in developers are not in a project to fully develop everything out the lead developer has to make hard choices unfortunately this normally end up in the majority case winning. With input device the majority case is users don't customise . Also the majority of users who do customise do have the skill to profile a device. That misses about 1 percent who are screwed. The lead developer is screwed by lack of resources so way to get extra resources is use the ones with the skills to profile devices at cost of 1 percent of the market and making 90%+ of users happy.

                Project under resourced does not make a nice project to deal with all the time.

                Comment


                • #38
                  Originally posted by oiaohm View Post
                  This is kind of a mistake. If you survey MS windows and Mac OS users 90%+ of them have never touched mouse/touchpad settings and don't even know where it is.


                  This is not true. The reality is majority of users don't care as long as its in the ball park its good enough they don't change the sensitivity to suite them.

                  Out the box experience is important for 90%+ of users. Yes you are in the less than 10% of users who care and want to adjust mouse and pointer setting at all.


                  How under resourced is a fact. This does effect how much the developer has to focus on market share and what is possible.


                  Yes for touchpads that libinput does not in fact have profile for this is required. You will notice quite a few users with particular brands say libinput is decent out the box because the vendor did the profiling for them but this is only 3 brands of laptops.


                  This is you miss reading the statement. There is not enough resources so now the lead maintainer has to optimise. Less complicated, Less bugs and suites 90% of users to be that way and fits with the amount of resources libinput has.

                  The reality you want more to handle more complicated with more work to maintain it you need more developers.

                  The reality is they don't want to do it arguement has you failing to step back look at the big picture see that you are in the minority and the developer is under resourced so has to make hard choices. Just like libinput does not come with a decent GUI program as part of libinput this is resources. Think about it what is the point about giving users mouse curve control if they don't have a good GUI to set it valves and diagnose if it right. Lack of GUI developer has knock on effects.

                  The reality here you are trying to get 1L of water out of a cup of water. If the resources as in developers are not in a project to fully develop everything out the lead developer has to make hard choices unfortunately this normally end up in the majority case winning. With input device the majority case is users don't customise . Also the majority of users who do customise do have the skill to profile a device. That misses about 1 percent who are screwed. The lead developer is screwed by lack of resources so way to get extra resources is use the ones with the skills to profile devices at cost of 1 percent of the market and making 90%+ of users happy.

                  Project under resourced does not make a nice project to deal with all the time.
                  This explains the state of Linux on the desktop!
                  This is kind of a mistake. If you survey MS windows and Mac OS users 90%+ of them have never touched mouse/touchpad settings and don't even know where it is.
                  Who says this?
                  Windows allows you to make changes, you don't have to be a computer guru to do it.
                  Nobody has the truth in their pocket, there are users who modify it, others who are comfortable with the default settings.

                  Comment


                  • #39
                    Originally posted by caligula View Post
                    Isn't SDDM pretty much dead
                    Yes, pretty much. https://github.com/sddm/sddm/pull/1367 and https://github.com/sddm/sddm/pull/1379 are open – the first being a KDE developer reviving old Wayland patches and the latter being an improvement over the first but in nine days nobody cared to review it. I don't get the attachment to it. SDDM has always been inferior to LightDM with the LightDM-KDE greeter which KDE used in the Plasma 4.x days. For a change Canonical wrote useful software and certain vocal people within KDE lobbied against it, despite the fact that greeters are completely autonomous projects that aren't being touched by LightDM's CLA at all. According to https://github.com/canonical/lightdm...addec1e081e5c7 LightDM supports Wayland greeters and sessions since six years. Lower commit numbers are not so important with LightDM because it's only the back-end technology and most work needs to go into greeters anyway.

                    I've had good experiences using GDM, btw. Runs under Wayland, under active development, packaged everywhere.

                    Comment


                    • #40
                      Originally posted by Charlie68 View Post
                      This explains the state of Linux on the desktop!
                      Being under resourced is a big problem.

                      Originally posted by Charlie68 View Post
                      Who says this?
                      Windows allows you to make changes, you don't have to be a computer guru to do it.
                      Nobody has the truth in their pocket, there are users who modify it, others who are comfortable with the default settings.
                      The reality here is hard. There are many surveys on what setting users change with windows. In university and business computer networks its important to know what setting changes you can in fact disable and users will not notice. The reality is 90% of users have to have a use able default settings because they are not going to change the mouse/touch pad settings.

                      Hard fact just because Windows or any other OS provides a feature does not mean many users in fact use it. So out the box defaults on touchpads, mice. keyboards... have to be generically decent as majority of users need that as that the only setting they will be using.

                      Windows does not allow you to change everything with input devices. The changes you make in Windows is already have limits placed on it by the device profile.

                      Yes like it or not people to complain about Windows not providing enough options either. So windows does not allow you to make enough changes to fix touchpad issues either. Yes this is because you are stuck behind the limitations of the device profile that you cannot change and will not be changed by the vendor at all. Remember libinput you submit data on issue with a device profile new versions to come out fixing it.

                      Charlie68 reality here Linux and Windows sux in different ways with input devices.
                      1) Windows if the device profile in a input driver is wrong the vendor is 99.9% of the time never going to fix it. Why that driver that came with windows computer most likely no longer has a developer so is a completely dead not update able item. libinput we are lucky here is a device profile issue is fixable where windows users are stuck with broken device profiles and end up buying a new computer to fix. Yes needing to submit a bug report to fix your device is better than the Windows answer with device profile problems of go buy a new computer because there is no developer here to process you complaints or fix it.
                      2) Window provides a few minor alteration options with GUI that are not enough to deal with major touch pad issues.
                      So Windows does not display solution either.

                      The reality here the proper fix requires someone who is a developer who can make decent GUI tools to join libinput. Then processed to take the diagnostic tools that libinput provides and make a GUI interface to profile the device including create custom profile. A tool like this has to exceed what Windows offers because the Windows offering presumes you have a base device maker profile to work on top of that on Linux you don't have in a lot of cases and the provided base profile with windows can be horrible wrong with no way to fix it.

                      Yes who is going to pay or be that GUI developer to fix this problem properly.

                      The reality lot of the stuff you are being asked to-do with libinput command line tools to fix a touch pad you have to-do so you have a profile of the device. GUI walking you though the process could in a few places could be done into mini games to make the process more fun and still collect the need data.

                      Charlie68 the problem here is you have been very tunnel visioned when you step back look at input devices from a big picture point of view it does not matter if you look at Windows, Mac, Linux, Android... Its all broken just in different ways. No party at this stage has the complete fix. The complete fix will take serous resources that libinput is missing and will require a profile device tool as in all that debuging tool crud of libinput presenting in a user friendly way.

                      Comment

                      Working...
                      X