Hacker News Reader

Comment by ZeroCool2u | Hacker News Reader

Damn. This is the first open source project I've ever sponsored and been really excited about sponsoring.

However, I've been using Ubuntu on Wayland for the past 2 years and have thoroughly enjoyed it. Frankly, I'm not willing to move back to X and it seems like a step backwards to me to invest any time, effort, or money in X at this point, so I'm going to discontinue my contribution.

I was hoping my monetary contribution would be a small ($25/Month) way to push Wayland forward, but the approach Bill and Povilas are taking is doing exactly the opposite by making it more comfortable for folks to stick with X. I understand there is disagreement and apprehension among the community with the slow move to Wayland, perhaps not at the same scale of systemd, but I firmly believe it's the best path forward and that X should be left behind.

View on HN

Povilas from the linked article here.

Having support in X (even if only in proposal stage) will allow much easier work in the toolkit and application layers. X and Wayland cover essentially al Linux users, so we will not need to estimate how many users will benefit from implementing gesture support in toolkit X or application Y. It will make convincing project maintainers easier.

This is really important, because maintainers of open source projects don't usually care about suggested features if they're not interested in them themselves. A new feature means additional work to them - discussing the design, reviewing PRs and handling of eventual bugs. Often the person who implements a feature disappears after the PR is merged. This makes the maintainers to view all feature proposals with a grain of salt. We need to have a convincing story of how majority end users would benefit in order to make the contributions easier.

Technical skill is not enough when contributing in open source. Politics are often as important.

Hi Povilas, thanks for the reply.

To be clear, you're saying that implementation in the applications/toolkits ecosystem is the barrier to a better trackpad experience on Linux and that by unifying the feature set of both X and Wayland, this will encourage that implementation among the developers of the applications/toolkits?

If that's correct, that is one angle I had not considered before. It's also somewhat unfortunate that it was not spelled out more clearly in Bill's post, but hindsight is 20-20 of course.

> To be clear, you're saying that implementation in the applications/toolkits ecosystem is the barrier to a better trackpad experience on Linux and that by unifying the feature set of both X and Wayland, this will encourage that implementation among the developers of the applications/toolkits?

Yes, exactly. AFAIK only Gtk right now supports any kind of touchpad gestures on Linux. Also, most of the users are still on Xorg and it's right now unknown when Wayland will be the default universally. So we would be in a weak position to demand maintainers attention, as there's little immediate benefit.

That's unfortunate. AFAIK GTK4 will be Wayland-only and out in a few months, and some bleeding edge distros such as Fedora already run Wayland out of the box, except if you have NVIDIA, because Xwayland is still crap under nv; but not for long, apparently there's some rumours and intention to improve the nv/xwayland situation, just nothing concrete out yet [1].

So yeah, betting on Xorg in 2020 is not a smart move IMO.

1: https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-G...

Distros keep trying to move to Wayland and then keep coming back to X11 due to usability. GTK4 is not going to help that.

> AFAIK GTK4 will be Wayland-only and out in a few months

I think this is incorrect. See the following, X11 backend is listed among supported backends. https://developer.gnome.org/gtk4/stable/extra-configuration-...

I share this sentiment and have also withdrawn my contribution. I remember reading X is in maintenance mode, and putting countless of hours into a plug-in for it because of some people’s refusal to switch to Wayland just seems futile and a waste of effort me, especially since, as they have mentioned, there already is a ‘happy path’. With the next LTS release of Ubuntu, both Fedora and Ubuntu will be Wayland by default. That’s what, something like 80% of all desktop distros in use?

In the end it is not my money so I cannot decide, but IMO they should focus on bringing Qt and WxWidgets up to snuff with GTK gesture support. That means that as long as people are on Wayland, gestures will work ‘everywhere’.

Qt even already has QNativeGestureEvent stuff, seems like it's only wired up on macOS for now.

Also, what really sucks about Qt is there's no kinetic scrolling in classic qt widget apps (e.g. telegram-desktop). Or there is but every app has to enable it or something??

Same here. I'm withdrawing my contribution if they're just going to short-term solutions which will lead toward slowing down the migration to Wayland. This is an irresponsible use of funds.

I think the strategy seems reasonable. It seems much easier for toolkits and application developers to feel like supporting gestures is worthwhile if they know that the large fraction of their users using X will get this feature. I think otherwise it can be easy to only want to implement things for the intersection of X and wayland.

Reply on news.ycombinator.com