Wayland on NetBSD – Trials and Tribulations | Hacker News Reader

Wayland on NetBSD – Trials and Tribulations

blog.netbsd.org

22 pointsjayp14188 comments

>In general, Wayland is moving away from the modularity, portability, and standardization of the X server.

This is the core criticism that's been advanced since a decade ago. X is the standard, it's an actual standard. Wayland is apparently both a protocol which is defined by an XML file, a reference implementation, but yet not a standard. It's going to be tough to get it into deployment until it is standardized and finished, so to speak.

It also doesn't give me any benefits, the programs I use all use X so Wayland becomes a parent process to X and that's apparently its job. Why not just skip it? X is very time worn and honed to a sharp edge, it can run on incredibly minimal hardware. Wayland depends on a lot more.

The criticism on Wayland usually tends to oversee that its developers are the same that used to work on X.

Several Linux distributions for embedded deployment, ChromeOS and WSL 2 (in the near future), Fedora, Ubuntu, all make use of Wayland

I don't think any of the Wayland developers worked on X before RedHat became an essential sponsor to X.org which is, in essence, finished software.

What's the function of the X.org maintainers now? Mainly merging drivers. It's been like this for over a decade, heck back in the old days several bold moves were made for network transparent 3D acceleration with X.org including working patches but it was never realized.

With everybody else going their own way, it might be time to revisit and reinvigorate these efforts.

Sure they did -- the old guard of Daniel Stone, Peter Hutterer, Adam Jackson, even Kristian Hoegsberg, they were all Xorg hackers before Wayland was a thing, and only a few had ever worked at Red Hat.

X.org doesn't really have maintainers, ever since Adam Jackson stepped away as release manager a few years ago. Alan Coopers it merges security patches, but most device driver support is handled in the lower levels now.

the core difference now is that Sun and tens of other unix vendors are dead so there's less competing commercial influences to the Linux crowd involved in maintaining this stuff.

note that Alan Coopersmith is paid by Oracle to work on Solaris.

> In pkgsrc we've patched the libraries to add kqueue(2) support, but the patches haven't been accepted upstream

How open are Wayland's developers to receiving patches to add support for platforms other than Linux?

"Everything is an extension" architecture is great when there's an incentive to promote extensions to some sort of core/standard library and then make them available everywhere.

If you look at the libxcb and xorgproto repositories you'll find something very similar to Wayland but for X (an RPC framework for building display servers / clients). The main difference is the X11 extensions are already known quantities thanks to X's monolithic history.

Developing a display server standard is a problem that Microsoft and Apple don't have to worry about.

Thankfully, hence why I was already running RDP sessions over WAN about 15 years ago, while X was chocking on similar connections.

Reply on news.ycombinator.com